The Internet of Things - Deutsche Bank [PDF]

Deutsche Bank Markets Research Industry

The Internet of Things

Date

6 May 2014 North America

United States TMT

Brian Modoff Research Analyst (+1) 415 617-4237 [email protected] Vijay Bhagavath, Ph.D Research Analyst (+1) 415 617-3324 [email protected] Kip Clifton, CFA Research Associate (+1) 415 617-4247 [email protected]

F.I.T.T. for investors Not quite the Jetsons yet, but places to look Internet of Things (IOT) is a big idea but promise and reality are a distance apart IOT is a set of multi-year secular growth themes in our view impacting every player in our universe. IOT has the potential for unlocking Trillions of dollars of value – through structural improvements in operational efficiencies in every industry sector. Early stage IOT initiatives in Smart Energy, Smart Retail, Cyber Security are the basis for our conviction on a long-tail positive impact to select players in our universe. But significant work lies ahead to make the promise a reality. It is our view that the ultimate goal of IOT is a ways off and may not have a significant impact on unit volumes of connected things until 2016. Still, we believe IOT is the biggest shift to hit our industry since the internet itself.

________________________________________________________________________________________________________________ Deutsche Bank Securities Inc. Deutsche Bank does and seeks to do business with companies covered in its research reports. Thus, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. DISCLOSURES AND ANALYST CERTIFICATIONS ARE LOCATED IN APPENDIX 1. MCI (P) 148/04/2014.

Deutsche Bank Markets Research North America

United States TMT

Industry

The Internet of Things

Date

6 May 2014

FITT Research Brian Modoff

Not quite the Jetsons yet, but places to look

Research Analyst (+1) 415 617-4237 [email protected]

Internet of Things (IOT) is a big idea but promise and reality are a distance apart IOT is a set of multi-year secular growth themes in our view impacting every player in our universe. IOT has the potential for unlocking Trillions of dollars of value – through structural improvements in operational efficiencies in every industry sector. Early stage IOT initiatives in Smart Energy, Smart Retail, Cyber Security are the basis for our conviction on a long-tail positive impact to select players in our universe. But significant work lies ahead to make the promise a reality. It is our view that the ultimate goal of IOT is a ways off and may not have a significant impact on unit volumes of connected things until 2016. Still, we believe IOT is the biggest shift to hit our industry since the internet itself.

Vijay Bhagavath, Ph.D

Near-term opportunities include We highlight three themes – Network Intelligence and Security, Edge Clouds and Datacenter Clouds, which form the basis for building a new generation of infrastructure for implementing “Internet of Things” use cases. IOT infrastructure rollouts are an incremental +$2B a year capex opportunity in our view (above and beyond the +$47B a year networking IT spend) and could be basis for upside to the CY16+ consensus view for select IT vendors. Primer and roadmap for investment This research note doubles as a primer on IOT. We walk through all elements of IOT from the architectures and protocols being considered to security and Big Data challenges. All of this is discussed with the beginner in mind, so that investors come away with a holistic view of the challenges that remain but also a sense of the possibilities both in the near term and in the long term. Key names Qualcomm: Qualcomm is likely one of the best positioned companies in the space, as it will have generated significant “lock-in” given it developed much of the foundational elements that drive even the most peripheral areas of IOT. Cisco: We have near-term caution on the company’s core business, while remaining constructive on CY16+ opportunities in IOT and product refresh cycles in datacenter switching, security and carrier routing. CommScope: The Wireless segment has the general trend of densification in its favor, but Enterprise segment could be impacted relatively more by IOT. F5 Networks: The market leader in Layer 4/7 is well-positioned to capture a meaningful percentage of new IT dollars that are being allocated for implementing the Network Intelligence and Security Iayers for IOT. Ciena: could benefit from new use cases around rollouts of optical Ethernet switches in IOT use cases; also optical network automation tools. Infoblox: It is likely to benefit from IT spending on IP Address Automation and DNS Security needed for implementing “Cloud-scale” Internet architectures. Other companies that could benefit from these trends include CAVM, JNPR, UBNT, ARUN, RKUS, ADNC and XXIA. We also include descriptions and exposure to over 50 other companies.

Research Analyst (+1) 415 617-3324 [email protected] Kip Clifton, CFA Research Associate (+1) 415 617-4247 [email protected] Top picks Qualcomm (QCOM.OQ),USD78.99 Commscope (COMM.OQ),USD27.84 Cisco Systems (CSCO.OQ),USD22.94 F5 Networks (FFIV.OQ),USD103.72

Buy Buy Hold Buy

Source: Deutsche Bank

Companies Featured Qualcomm (QCOM.OQ),USD78.99 Buy 2013A 2014E 2015E EPS (USD) 3.91 4.56 5.26 P/E (x) 16.4 17.3 15.0 EV/EBITDA (x) 3.3 3.9 2.9 Commscope (COMM.OQ),USD27.84 Buy 2013A 2014E 2015E EPS (USD) 0.12 1.43 1.51 P/E (x) 138.9 19.5 18.4 EV/EBITDA (x) 8.2 8.5 7.9 Cisco Systems (CSCO.OQ),USD22.94 Hold 2013A 2014E 2015E EPS (USD) 1.86 1.42 1.65 P/E (x) 11.0 16.2 13.9 EV/EBITDA (x) 5.5 7.4 6.2 F5 Networks (FFIV.OQ),USD103.72 Buy 2013A 2014E 2015E EPS (USD) 3.50 3.90 4.87 P/E (x) 25.1 26.6 21.3 EV/EBITDA (x) 12.0 12.6 9.4 Source: Deutsche Bank

Valuation and risks The stocks in our wireless and data networking peer group trade at 17-18x forward P/E, a premium to the market, given the above-market growth prospects of the networking universe. Key risks (downside and upside) are from unanticipated shifts in IT spending, share shifts and technology disruptions. ________________________________________________________________________________________________________________ Deutsche Bank Securities Inc. Deutsche Bank does and seeks to do business with companies covered in its research reports. Thus, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. DISCLOSURES AND ANALYST CERTIFICATIONS ARE LOCATED IN APPENDIX 1. MCI (P) 148/04/2014.

6 May 2014 TMT The Internet of Things

Table Of Contents

Executive Summary............................................................. 4 Companies: ........................................................................................................... 7

IOT Vision .......................................................................... 11 Key Building Blocks of an Internet of Things Infrastructure ........................................................................................... 15 Architecture ....................................................................... 22 IOT Protocols ..................................................................... 28 No clear protocol in place .................................................................................. 29

Security ............................................................................. 37 APIs ................................................................................... 40 Middleware ......................................................................................................... 45

IOT as a Platform ............................................................... 47 IOT platform for devices ..................................................................................... 50

Are Networks Ready? ........................................................ 54 The rising Intelligence in Sensor Nodes ............................ 58 Frequency challenges ........................................................ 61 Big Data and Analytics ...................................................... 64 IOT Use Cases ................................................................... 69 Companies......................................................................... 81 Qualcomm .......................................................................................................... 81 Cisco ................................................................................................................... 83 CommScope ....................................................................................................... 84 F5 ........................................................................................................................ 84 Ciena ................................................................................................................... 86 Infoblox ............................................................................................................... 86 Juniper ................................................................................................................ 87 Riverbed .............................................................................................................. 87 Cavium ................................................................................................................ 87 Ubiquiti ............................................................................................................... 88 Aruba and Ruckus .............................................................................................. 88 Audience ............................................................................................................. 89 Ixia ...................................................................................................................... 89 Companies outside of our coverage area .......................................................... 89

Page 2

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Table of Exhibits Figure 1: Hype versus Reality ............................................................................................. 5 Figure 2: IOT unit forecasts ................................................................................................ 5 Figure 3: Current IOT use cases ......................................................................................... 5 Figure 4: Places to invest ................................................................................................... 5 Figure 5: Internet of Things – A Big Picture View ............................................................. 12 Figure 6: Potential operational efficiency savings from representative Internet of Things industrial use cases (based on GE IOT Study Data) .......................................................... 13 Figure 7: Contrasting Legacy Datacenter and IOT Edge Cloud / Datacenter Cloud architectural approaches .................................................................................................. 15 Figure 8: Key Edge Cloud feature-functions (e.g. Cisco ISR 819 series, OEM enterprise edge routers, etc.) ............................................................................................................ 16 Figure 9: Key Datacenter Cloud infrastructure layers ....................................................... 19 Figure 10: Conceptual view of an IOT Datacenter Cloud – illustrating the Information Technology (IT) and Operational Technology (OT) infrastructure elements...................... 21 Figure 11: Layers of a generic oneM2M architecture....................................................... 23 Figure 12: Architecture proposed by IOT-A ...................................................................... 24 Figure 13: WSNs market traction as sensor costs decrease ............................................ 25 Figure 14: Tracy and Sreenan architecture....................................................................... 27 Figure 15: Managing disparate devices/services and transport layers ............................. 28 Figure 16: Members of AllSeen Alliance .......................................................................... 30 Figure 17: MQTT machine-to-machine communication ................................................... 32 Figure 18: Abstract layering of CoAP ............................................................................... 33 Figure 19: Architecture of a CoAP-based wireless sensor network.................................. 33 Figure 20: 6LowPAN enables IPv6 over low power WSNs .............................................. 35 Figure 21: Google maps via Verizon’s website ................................................................. 41 Figure 22: Directions to store via Google maps ............................................................... 41 Figure 23: Mashups make it relatively easy to combine applications .............................. 42 Figure 24: Sensor middleware, gateway and API management....................................... 44 Figure 25: Enabling interoperability via APIs .................................................................... 44 Figure 26: Gartner Magic Quadrant .................................................................................. 45 Figure 27: ThingWorx Application development platform................................................ 49 Figure 28: Qualcomm’s IOT development platform ......................................................... 51 Figure 29: Broadcom WICED module .............................................................................. 52 Figure 30: Arduino microcontroller board ........................................................................ 53 Figure 31: Galileo board ................................................................................................... 53 Figure 32: IOT brings value to businesses, consumers and assets .................................. 54 Figure 33: Ericsson Device Connection Platform – general deployment .......................... 55 Figure 34: Gateways in the context of IOT ....................................................................... 56 Figure 35: DRX cycle effect on battery life ....................................................................... 57 Figure 36: A partial list of sensor use cases ..................................................................... 59 Figure 37: Radio technology range and bit-rate ............................................................... 62 Figure 38: 2013 Worldwide Big Data Revenue by Vendor ($ mlns).................................. 64 Figure 39: How IOT will effect the Big Data landscape .................................................... 65 Figure 40: Stream processing enables real-time response ............................................... 66 Figure 41: iTRACS DCIM to optimize data center for stream processing ......................... 67 Figure 42: The path to increased investment in DCIM ..................................................... 68 Figure 43: Industry data summary on CY15 network equipment capex spending [Infonetics] and DB view of incremental network infrastructure spending on IOT use cases ................................................................................................................................ 69 Figure 44: Illustration of IOT Cyber Security network architecture .................................. 71 Figure 45: Illustration of key IT stacks involved in implementing an IOT Cyber Security architecture ...................................................................................................................... 74 Figure 46: Conceptual view of an IOT architecture to support a Smart Energy use case. 75 Figure 47: Conceptual view of relevant IT stacks involved in the Smart Energy use case 76 Figure 48: Conceptual view of an IOT architecture to support a Smart Retail use case ... 79 Figure 49: Conceptual view of relevant IT stacks involved in Smart Retail use case ....... 80 Figure 50: Qualcomm’s IOT initiatives.............................................................................. 81 Figure 51: GSMA unit estimates for the connected car ................................................... 81 Figure 52: Unit cellular estimates for the connected home .............................................. 82

We wish to thank Vinod Khurana and Mohd-Tabish Shams of Evalueserve for their contributions to this work

Deutsche Bank Securities Inc.

Page 3

6 May 2014 TMT The Internet of Things

Executive Summary "A global network infrastructure, linking physical and virtual objects through the exploitation of data capture and communication capabilities. This infrastructure includes existing and involving Internet and network developments. It will offer specific object-identification, sensor and connection capability as the basis for the development of independent cooperative services and applications. These will be characterized by a high degree of autonomous data capture, event transfer, network connectivity and interoperability." – The CASAGRAS report 2011 The above quotation is one for the more succinct descriptions of what the Internet of Things (IOT) is proposed to be. As to how this translates into units and dollars, ideas begin to differ among third-party data providers, industry bodies, companies and people. While almost everyone agrees that the benefits will be significant, the target beneficiaries and when they are affected is a constant debate among pundits. The most noise is made by companies themselves, proclaiming that IOT will have a huge impact to their bottom lines and that this impact is right around the corner. Our thesis differs from most. We believe that IOT, as it is described in the ideal sense above, is a ways off, and may not have a significant impact on unit volumes until 2016 at the earliest. There are a number of challenges around architecture, protocols and security, all of which need to be addressed before IOT reaches the tipping point of critical mass adoption. However, despite this time lag, we believe there are areas of incremental growth which can be exploited now as companies are winning in these “smart silos,” places like the connected car, smart home and the smart grid. We walk through the impact of these investments, but only as it relates to the companies in wireless and data networking universe. This research note doubles as a primer on IOT. We walk through all elements of IOT from the architectures and protocols being considered to security and big data challenges. All of this is discussed with the beginner in mind, so that investors come away with a holistic view of the challenges that remain but also a sense of the possibilities both in the near term and in the future. It is within this framework that we conclude the note with a company exposure list and highlight those names which we believe have the best exposure, and those with a foothold into the likely winning platforms of the future.

Our thesis differs from most: We believe the IOT, as it is described in the ideal sense above, is a ways off, likely not having a significant impact on unit volumes until 2016 at the earliest

There are a number of challenges around architecture, protocols and security, all of which need to be addressed before IOT reaches the tipping point of critical mass adoption

Where we stand today We are in the midst of a decently sized hype cycle. In many ways this is positive, as it attracts developers and stimulates idea generation. In other ways, it can lead people to get ahead of themselves in terms of unit volumes in the near term. A quick glance at some of the unit forecasts gives a decent sense of how quickly some project the unit volume to grow. There are caveats around every estimate, some around the definition of units (e.g. is a jet engine one IOT unit or hundreds?) and others around a decision tree of actions (e.g. one protocol used) working in just the right way to achieve targets. In spite of these caveats, our research on the unit estimates suggests that we are likely further out than most project given the challenges.

Page 4

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 1: Hype versus Reality Present

Figure 2: IOT unit forecasts 26 billion by 2020 Gartner

Hype cycle

50 billion by 2020 Cisco

Hype/Revenue

15 billion by 2015 Intel 50 billion by 2020 Ericsson

IOT Unit Sales

18 billion M2M devices by 2022 Machina Research Source: Various sources – see above

2017 – 2020 timeframe

Time Source: Deutsche Bank

If we place ourselves at the dotted line, unit volumes are still a ways off and we are in the midst of a considerable hype cycle. Still, we believe there are places to invest, as there are companies seeing meaningful sales within the realm of IOT. These are often areas where the environment is contained, disparate protocols need not be leveraged, the overall architecture has been decided upon and security is less of an issue; targeted industries like the ones mentioned to the right, with some having a greater traction than others (e.g. fitness having greater traction than medical given privacy issues), are the value creators within the current IOT. Eventually every industry will leverage IOT to realize efficiencies, and this will have a profound effect not just on unit sales, but services as well. Our goal with this paper is not to focus on the dreams of what could be but rather on what can be done with the current tool sets and provide a framework to help investors find companies within the circle below.

Still, we believe there are places to invest… …these are often areas where the environment is contained, disparate protocols need not be leveraged, the overall architecture has been decided upon and security is less of an issue

Figure 4: Places to invest Present Hype cycle

Figure 3: Current IOT use cases Shipping and location

Hype/Revenue

Home automation

IOT unit Sales

The smart grid Farming/agriculture Medical Fitness

Companies are winning here, in various “smart silos” Time Source: Deutsche Bank

2017 – 2020 timeframe

Retail inventory management Environment monitoring Industrial automation Security (building and infrastructure) Vehicle/smart car Lighting Asset tracking Source: Deutsche Bank;

Deutsche Bank Securities Inc.

Page 5

6 May 2014 TMT The Internet of Things

Areas such as asset tracking, shipping, home automation and smart cars leverage the current wireless infrastructure and will increase their reliance on such connections in the future. Even though most of the “things” we think about are connecting to the IOT using close connections over unlicensed – such as Bluetooth, WLAN, NFC – the data from these “things” will eventually flow back to the network, where they will be analyzed and managed. This means that the connections themselves, whether wireless or wired, need to be ready for this data deluge and we believe that most networks around the world are not to ready to support this sort of traffic in a reliable fashion. Potentially trillions at stake, but over the long term IOT is a multi-year IT spending and infrastructure buildout initiative involving the private sector and governments and is best viewed as a trillions of dollars opportunity in the long term created from rich interconnections among people, processes, data and things. While IOT is an evolving opportunity, our IT conversations suggest a potential for $2B in incremental annual capex spending over the next several years on Layer 0/7 network and wireless equipment (e.g. switches, routers, optical transport, security, Layer 4/7 appliances) that would be required to implement IOT use cases in these smart silos.

We note potential for $2B in incremental annual capex spending over the next few

We view the +$2B in incremental capex spending on IOT use cases – above and beyond the current +$47B a year in IT spending on Layer 0/7 wireless and network equipment – as a fundamental basis for potential upside to the CY16+ consensus view for select IOT-exposed stocks in our networking universe. We believe the basis for a conservative $2B a year in incremental IT spending on IOT use cases is that CxOs would likely allocate around 10% of the $18B a year in opex savings from implementing IOT. (A recent GE IOT study notes a potential for +$276B in opex savings over the next 15 years from a modest 1% improvement in key operational efficiency metrics from leveraging IOT.)

years on Layer 0/7 network equipment. The +$2B in incremental capex spending would be above and beyond the current +$47B a year in IT spending on Layer 0/7 network and wireless equipment.

However, challenges still remain. We walk through the challenges associated with every element of IOT from the architecture down to Big Data and analytics. In a sense, IOT is different from every other emerging technology. With the Internet, there was one protocol and developers were able to build to that. With cellular technology, a standard body now helps to oversee where the standard goes and everyone builds to that. IOT does not yet have such a body and every smart silo is building in a different direction, often with little thought to how integration efforts might go. Given this difference, we discuss where the standards might go, which platforms have the early foothold and what architecture seems most appealing.

In a sense, IOT is different

Our investment thesis centers on names that are being impacted now and will see this impact accelerate along with the adoption of IOT. On the Wireless side: Qualcomm, CommScope, Audience, Aruba, Ruckus and Ubiquiti. On the Data Networking side: Cisco, F5 Networks, Ciena, Cavium, Juniper, Infoblox, Ixia and Riverbed.

Our investment thesis centers

Page 6

from every other, emerging technology. With the Internet, there was one protocol and many were able to build to that.

on names that are being impacted now and will see this impact accelerate with the adoption of IOT.

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Companies: Qualcomm Qualcomm has been positioning IOT as the growth driver for the company since cellular wireless technology first gained traction. The company started with Omnitracs, a platform for truck fleet management, and then branched out into a number of IOT initiatives. It developed the AllJoyn protocol, which is now overseen by the Linux foundation and has more partners and contributors than any other protocol platform being considered. Qualcomm will be a key beneficiary of IOT, which is not an extraordinary statement. But given our view that it will take time for the market to develop, and with a company that is on the cusp of shipping 1 billion ASICs per year, it will be some time before it is significant for the company. Having said that, when IOT unit volumes begin to ramp up in earnest, Qualcomm will likely be one of the best positioned companies in the space, as it will have generated significant “lock-in” given it has developed much of the foundational elements which drive even the most peripheral areas of IOT. Cisco The basis for our constructive view on Cisco in the IOT infrastructure opportunity in the out quarters (2H15+) is: 1. Cisco is viewed as a thought leader in IOT (we note CEO keynotes on IOT at major trade shows) in IT circles and has an IOT focused cross business unit product team. 2. Cisco has the broadest set of touch points in enterprise and service provider infrastructure and a broad portfolio of IT services from campus to the core. 3. Cisco’s IOT opportunity in 2H15+, given our view that the company has the set up, from a product portfolio, customer base, market share and sales execution POV, for offering a packaged bundle of IOT solutions involving millions of smart objects (i.e. Internet connected wireless sensors) connecting to an Edge Cloud. 4. The stock could be a potential margin improvement and earnings growth story in CY15+, if the company gradually exits from several slower growth and focuses on higher margin “software” and “advanced services” opportunities in Private Clouds, Telco Clouds and Internet of Things (refer to our recent CSCO Growth Strategy note. Our Hold rating on Cisco reflects our near-term caution on the company’s core business (switching and routing in product transition mode), while remaining constructive on CY16+ opportunities in IOT and product refresh cycles in datacenter switching, security, carrier routing, among others. CommScope The Wireless business segment has been a key driver for CommScope’s overall top-line. While there is nothing specific in Wireless that falls within IOT per se, its DAS solution will be a key enabler of IOT, as it puts the cellular network closer to the user. The company has made some interesting acquisitions in the Enterprise segment, mainly within the vein of IOT – Redwood Systems, an intelligent lighting company and iTRACs, a data center infrastructure management platform. The two companies were acquired when CommScope was held privately. We discuss both of these, as well as other segments in our report. In short, the trends in IOT will impact Wireless and Enterprise segments. While Wireless has the general trend of densification working in its favor, it is likely the Enterprise segment will be impacted relatively more by IOT.

Deutsche Bank Securities Inc.

Page 7

6 May 2014 TMT The Internet of Things

F5 Networks Buy-rated F5 is our top midcap idea to play the near-term [CY15/16+] spending on IOT – given our research noting that Network Intelligence and Security are a Phase 1 priority in the multi-year rollout of IOT use cases in enterprise and service provider verticals. A third of the +$2B in incremental networking capex a year (DB view) on IOT use cases could be utilized for Network Intelligence and Security solutions (we note: Layer 4/7 capex is typically a third of datacenter networking capex). This is basis for our view that the Layer 4/7 category leaders (FFIV; BLOX, etc) could see modest TAM expansion in the $100s of Millions a year range from enterprise IOT use cases such as Industrial Cyber Security, Smart Energy, Smart Retail, etc and also service provider IOT use cases such as e-Healthcare, home automation, etc in the out quarters. Ciena Buy-rated Ciena could benefit from sales of higher margin optical Ethernet access products and optical network automation software tools to support IOT Edge Cloud buildouts. We also note opportunities for Ciena’s Carrier Ethernet solution and optical network management software tools in the Smart Energy use case. Undersea cable – carrying optical Ethernet data traffic – is a relevant use case for sales of Ciena’s carrier Ethernet switches at either ends of the offshore and onshore optical WAN links. Ciena’s optical network management SW could be used to dynamically provision optical network transport equipment at either ends of an optical fiber link. Infoblox Infoblox – a market leader in DNS and IP Address Management [IPAM] solutions – could secularly benefit from the IT spending on IP Address Automation and DNS Security – needed for implementing “Cloud-scale” Internet architectures in the Edge Cloud and Datacenter Cloud – two key themes of our FITT. We have a long view on Infoblox – noting that the company’s IP Address Automation (DNS, DHCP, IP Address Management; collectively IPAM) opportunity could meaningfully scale from – the current +25M enterprise application server IPAM opportunity [refer to our BLOX coverage initiation note for further color] to potentially offering IPAM capabilities for 100s of Millions of Internet Connected Objects in key enterprise and service provider IOT use cases. Juniper We are cautious on Hold-rated Juniper in IOT networking buildouts. Juniper, in our view, lacks a competitive best-in-class Enterprise Edge Cloud and Datacenter Leaf and Spine Cloud switching portfolio. We see a limited role for Juniper in the access segments of IOT use cases, i.e. in WiFi, campus switching or in campus and branch security. Our cautious view is based on Juniper’s low single-digit market share in enterprise networking and lack of meaningful competitive advantage versus market leader Cisco across a wide range of IOT enterprise networking segments. Riverbed Hold-rated Riverbed could modestly benefit from enterprise IOT use cases in CY15/16+, primarily in network performance monitoring by leveraging Riverbed’s Cascade product portfolio. That said, we remain cautious on

Page 8

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Riverbed, noting +70% of Riverbed’s FY revs are exposed to a declining growth WAN optimization (Steelhead) portfolio. Cavium While current levels suggest a neutral risk/reward equation – basis for our Hold rating – given, stock trading at ~25x P/E for ~20% consensus CY15 rev growth expectation, we have a constructive longer-term view on Cavium’s new multicore processors: specifically, Octeon3, Liquid IO and Neuron TCAM chips – which are relevant for IOT enterprise and service provider use cases. In the Datacenter Cloud, Cavium’s Liquid IO application acceleration processors could benefit from IOT use cases such as Big Data Analytics, processing high volumes of aggregated and encrypted IOT sensor data (SSL Offload, data compression, etc) – as HW acceleration add-ons in datacenter servers, etc. Similarly, Cavium’s Nitrox processors fit into HW acceleration use cases in Layer 4/7 ADC and Next-Gen Security appliances. Ubiquiti Ubiquiti is able to proliferate a number of markets quickly, given its low-cost and global network of distributors. Most investors know the company for its airMax and Unifi product lines; however, it has three other platforms – airVision, mFi and EdgeMax. mFi is Ubqiuiti’s foray into machine-to-machine communication. The mFi platform includes everything from smart power strips to sensors (motion, light, temperature and doors). While we believe the devices to have meaningful utility, their channel and end-customer (as it stands today) may not create the necessary demand to make this platform a significant contributor to the top line. However, if consumers could install these devices easily and that they were available in known retail locations, we believe that the traction would be much more significant. As it stands, we see it more as a hobby of a WISP operator or other tech savvy people.

While we believe the devices to have meaningful utility, their channel and endcustomer (as it stands today) may not create the necessary demand to make this platform a significant contributor to the top line.

Aruba and Ruckus We grouped Ruckus and Aruba as there is no product or platform directed at IOT by either company; however, both companies should see a general uplift from the broad trend of connectivity and other trends such as 802.11ac adoption, location analytics, BYOD, Hotspot 2.0 and cloud managed WLAN services. These trends will play a major role in making a wireless connection faster. Additionally, some IOT modules will connect via WiFi and require a dynamic usage of the spectrum – it will be up to the AP to understand this. This is where Aruba and Ruckus could differentiate their wares as both have distanced themselves by incrementally adding intelligence into their APs, and going further to integrate IOT elements would not be a reach for either company. In short, we believe that Ruckus and Aruba will eventually be beneficiaries of IOT, but it will likely not happen in the near term. Audience Audience is known for its audio and voice processors. But it could make an inroad into IOT with its Voice Q and Motion Q technologies that are based on voice and motion recognition, helping devices to wake up and sleep, and thereby efficiently managing battery life. The company recently released the demo versions of the processor at the Consumer Electronics Show and the Mobile World Congress. The processor will be integrated into the company’s eS700 series, which will be shipped in the second half of this year. While the processor may gain initial traction in smartphones, it could also be used in remote controls, wearables, fitness and health devices, mobile computers and Deutsche Bank Securities Inc.

Page 9

6 May 2014 TMT The Internet of Things

other battery powered devices. Going back to our original thesis, while we believe that it will be some time before this type of processing capability sees significant growth, we believe that it will be eventually included in adjacent markets and Audience could see meaningful upside. Ixia Hold-rated Ixia will modestly benefit in enterprise and carrier IOT use cases in the out quarters, in the areas of network visibility and Layer 2/7 network equipment port monitoring, leveraging its Anue acquisition.

Page 10

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

IOT Vision We view the Internet of Things as much more than a 2.0 remake of the Internet. In our view, the accurate definition of Internet of Things is a “networked” interconnection of people, business processes, data and things. In contrast, the Internet is mainly an interconnection of computing devices and applications.

Internet of Things is a “networked” interconnection of people, business processes, data and things

The fundamental difference between an Internet of Things infrastructure and a conventional IT infrastructure is the need for a new network element which we term an “Edge Cloud Node”. An Edge Cloud Node is an important element between the Datacenter Cloud and the IOT end-points for the following reasons (discussed in detail in the IOT Building Blocks section): 1.

Offering “localized” compute, networking, security, and storage feature-functions in physical proximity to the Internet-connected objects – so as to intelligently process raw data output from 1000s of IOT sensors and provide immediate local control decisions back to the IOT elements;

2.

Aggregate huge volumes of IOT data and send it upwards to centralized Datacenter Clouds – for Big Data Analytics, building a global view for centralized planning and decision making, etc.

3.

While conventional datacenters operate under the assumption of a low latency and continuously available network connection between the datacenter applications and the end-devices, the Edge Cloud, being application and network aware – is able to intelligently process IOT data (locally, when necessary) and trickle charge IOT data up to centralized Clouds – for business planning, generating longer-term operational efficiencies, etc.

Internet of Things use cases span a broad range of industry sectors from enterprise opportunities in smart manufacturing, energy, transportation and retail to public sector programs in smart traffic and city services and massmarket initiatives in e-health and home automation.

We note +$19 Trillion of IOT economic value creation

In Figure 5, we note +$19 Trillion of economic value creation potential and +25 Billion Internet connected devices, according to a recent Cisco IOT study.

Deutsche Bank Securities Inc.

potential and +25 Billion Internet connected devices

Page 11

6 May 2014 TMT The Internet of Things

Figure 5: Internet of Things – A Big Picture View

Internet of Things [IOT] +$19T Economic Value Creation Potential (CY13-22)

Private Sector

Public Sector

+$14.4T Value at Stake ‰ Smart Manufacturing (27% of value) ‰ Smart Retail (11%) ‰ Information Services (9%) ‰ Financial Services (9%) ‰ Healthcare, Prof. Services, Business Management, Education, Wholesale, etc (~44%)

+$4.6T Value at Stake ‰ Smart Connected Cities ‰ Opex Efficiencies in Govt Agencies ‰ Connected Defense ‰ Citizen Experiences ‰ Employee Productivity Improvement

Source: Deutsche Bank and Cisco IOT Study Data [bottom-up analysis of 21 private sector and 40 public sector use cases; www.cisco.com

Our vision of IOT is not an abstract futuristic vision. Instead, our thematic view of IOT is grounded in the economic value creation potential of Internet of Things which in our view drives the capex and people investment in Fortune 500 enterprises for implementing mainstream IOT use cases in healthcare, connected transportation, oil and gas, smart manufacturing, smart retail, etc.

Our view of IOT is grounded

A case in point is a recent GE IOT research study which notes that a modest 1% improvement in operational efficiency metrics in key industries such as aviation, energy, transportation and logistics and healthcare could drive a potential $276B in opex savings over the next 15 years (see Figure 6).

investment for implementing

Page 12

in the economic value creation potential of Internet of Things – which in our view drives the capex and people mainstream IOT use cases

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 6: Potential operational efficiency savings from representative Internet of Things industrial use cases (based on GE IOT Study Data)

Industry Vertical (IOT Use Case)

Operational Efficiency Improvement

Potential Opex Savings +$276B (15 yr timeframe)

Smart Energy (Oil and Gas)

1% reduction in capital expenditures

+$90B

1% fuel savings

+$66B

1% improvement in operational efficiency

+$63B

1% fuel savings

+$30B

1% improvement in operational efficiency

+$27B

Smart Utilities

Smart Healthcare

Smart Aviation

Connected Transportation

Source: Deutsche Bank and GE IOT Study Data

At a bigger picture level, a recent Cisco IOT study notes +$19 Trillion in value at stake that could be created from public and private sector enterprises leveraging Internet based interconnections among people, processes, data, and physical objects – i.e. the “Internet of Things” – to generate structural improvements in productivity, cost savings and user experience. The $19 Trillion in IOT value creation potential is based on a bottom-up economic analysis of +60 use cases and represents a 10-year NPV (CY13-22) of the potential economic benefits that could be generated in the private sector ($14.4T) and the public sector ($4.6T) from IOT solutions, summarized as follows:

Deutsche Bank Securities Inc.

Page 13

6 May 2014 TMT The Internet of Things

1.

Asset Utilization ($2-3T range) – generated from structurally lower operational costs through business process and capital utilization efficiencies;

2.

Employee Productivity ($2-3T range) – generated from efficiencies that result from higher productivity person-hours;

3.

Supply Chain and Logistics ($2-3T range) – based on improvement in business process efficiencies across the global supply chain;

4.

Customer Experience ($3-4T range) – based on increasing a customer’s lifetime value and TAM expansion through new customer additions;

5.

Reduced Time to Market ($3T range) – IOT improves the return on R&D investments and key business performance metrics such as Time to Market advantage, in addition to creating new revenue streams from new IOT business models and use cases;

6.

Public Sector ($4-6T range) – IOT helps in structurally lowering operating costs in government agencies and at the local/state/federal levels, e.g. in initiatives such as smart cities, smart connected buildings, supply chain and procurement optimization, digital government initiatives, etc.

Page 14

labor

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Key Building Blocks of an Internet of Things Infrastructure The foundational basis for Internet of Things is ubiquitous access for Internetconnected objects to IT capabilities such as computing, security, performance optimization, storage and wireless sensors via an “Edge Cloud” layer and a “Datacenter Cloud layer” for implementing the IOT use cases, as we summarize below:

IOT use cases need localized and low latency access to compute, networking, security and storage

Edge Cloud Internet of Things architectures are fundamentally different from currentgeneration Internet architectures from the point of view that IOT use cases need localized and low latency access to compute, networking, security and storage resources at the periphery of the IOT architecture (see Figure 7).

resources – hence the need for an Edge Cloud - at the periphery of the IOT architecture

Figure 7: Contrasting Legacy Datacenter and IOT Edge Cloud / Datacenter Cloud architectural approaches

Legacy Datacenter (Client/Server Applications)

Limited Scale Layer 4/7 Services (WAN Opt; ADC)

IT Datacenter

Datacenter Cloud (Web 2.0 / SaaS)

Datacenter Cloud

Layer 3 Leaf / Spine Switching Network & Virtual Overlays (Cisco Nexus 9k, bare-metal switches, etc)

Layer 2/3 Switching Network

Network Firewalling; DNS

App Specific QoS; Big Data Analytics; Traffic & Policy Management (F5, Cisco ACI, VMware, IBM…)

Application + Network Security; DNS (F5, Infoblox, Cisco, etc)

Assumes Variable-Latency; Variable BW Network Link Assumes Low-Latency; Constant BW Network Link

Edge Cloud Enterprise Edge Router w/ Local Compute, Networking, Storage, Security (e.g. Cisco ISR 819; HP MSR; OEM Edge Routers)

Client Devices Wireline & Mobile Devices PCs, Smartphones, etc

IOT End-Points Wireline & Mobile IOT End-Points; IOT Sensors, Actuators, Mobile Devices, etc

Source: Deutsche Bank

Deutsche Bank Securities Inc.

Page 15

6 May 2014 TMT The Internet of Things

In comparison, most current-generation Internet end-points application and compute resources from a Datacenter Cloud.

leverage

As a side comment, Internet round-trip delays in the 10s of millisecond timescales are adequate for most Web based and SaaS applications given our view that Web applications leverage layer 4/7 features such as client-side data caching and Web acceleration manage around Internet round-trip delays. Localized real-time access to Cloud computing resources is therefore not a requirement to run most Web-based and SaaS applications. We refer to the local compute, networking and storage resources as “Edge Cloud Nodes”, and to the network of Edge Cloud nodes as an “Edge Cloud Layer”. An Edge Cloud Node provides a rich set of localized compute, networking, security and storage feature-functions for Internet-connected physical objects – at millisecond-scale processing time – via a combination of wired and wireless access methods, e.g. 3G, 4G/LTE, WiFi, wireline Ethernet, etc. (see Figure 8). The Edge Cloud node also aggregates sensor data from multiple Internet connected physical objects for post processing in a Datacenter Cloud.

The Edge Cloud node aggregates data from multiple Internet connected physical

Figure 8: Key Edge Cloud feature-functions (e.g. Cisco ISR 819 series, OEM

objects for post processing in

enterprise edge routers, etc.)

a Datacenter Cloud

• Wireline Ethernet; 3G/4G; WLAN • Application Specific Network Bandwidth & Latency Provisioning (e.g. Cisco ACI Enterprise) • Automated Edge Cloud Platform Provisioning • Local Compute (e.g. Cisco UCS running Application VMs) • Local Flash Storage • Layer 2/3 Switching and Routing • IOT Sensor Data – Traffic Aggregation • Network Intelligence and Security (WAN Optimization; Application Optimization; Malware Defense; Access Control, etc) • Granular Traffic Analytics

Source: Deutsche Bank

Page 16

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Aggregating IOT sensor data back to a centralized Datacenter Cloud is useful for correlating IOT data sets across thousands of sensors and updating IOT objects with network security policies (e.g. access control rights, data encryption, etc.) and relevant IT and business rules. Cisco’s 819 Integrated Services Router (ISR) is an example of a commercially available Edge Cloud Node offered in industrial hardened and standard platform form-factors and supports 3G, 4G/LTE, Enterprise and Carrier WiFi, Ethernet LAN access and Layer 2/3 switching and routing capabilities. While Cisco is the market leader in enterprise routing with +75% market share Infonetics), our research suggests that the primary competition for Cisco’s ISR 819 router in IOT Edge Cloud deployments is likely to be from HP (MSR series), especially for the enterprise IOT use cases. Cisco’s ISR 819 router (and the branded enterprise routers, in general) is also likely to see growing competition from unbranded ODM enterprise routers which are customized Edge Cloud Nodes, optimized for a specific IOT use case and built using merchant silicon to implement lower-end (1G or lower) layer 2/3 switching, routing and network management features, in addition to the best in class wireline + wireless access. Note that the Cisco ISR Edge Cloud Node (and other branded vendor products such as HP’s MSR node) can support IOT objects that are in motion or nomadic within a local area with telemetry data from the sensors aggregated and processed via 4G/LTE or WiFi access. Cisco’s ISR node also supports Mobile IP protocols to dynamically build and manage IP routed networks among wireless sensor objects. As we head into CY15+, when the early set of IOT use cases in enterprise and service provider verticals would start scaling to 1000s of IOT objects, we see the role for an interconnection of Edge Cloud Nodes which we term as the Edge Cloud Layer. The Edge Cloud Layer has a meaningful value proposition in scale-out IOT use cases involving tens of 1000s of Internet connected objects connected to 1000s of Edge Cloud Nodes. For example: we see the Edge Cloud Layer ensuring security and IT policy coordination across multiple Edge Cloud Nodes (i.e. enterprise edge routers). The Edge Cloud Layer is ideal for ensuring that security patches and IT rules are applied consistently across Edge Cloud Nodes in order to avoid configuration mismatches among Edge Cloud Nodes (which are often the root cause of network outages, security breaches and performance issues). A related value proposition of the Edge Cloud Layer is to leverage the interconnectivity among Edge Cloud Nodes for utilizing spare compute and storage resources across enterprise edge routers (for example) to handle overload situations at any given edge routing node especially for objects that output Gigabytes of data such as Oil and Gas IOT sensors. In the IOT Themes chapter, we discuss a key set of IOT use cases that leverage the Edge Cloud Node and a networked interconnection of Edge Cloud Nodes to implement low-latency and localized compute, networking and storage capabilities. Deutsche Bank Securities Inc.

Page 17

6 May 2014 TMT The Internet of Things

Datacenter Cloud; Network Intelligence and Security While the Edge Cloud Nodes provide localized and real-time access to compute, networking and storage capabilities for IOT end-points such as wireless sensors and various other Internet connected objects, a Datacenter Cloud Layer could be viewed as a centralized Cloud layer that offers key features such as: 1.

Business Intelligence – based on Big Data Analytics of the information collected and processed across 1000s of IOT end-points;

2.

Cloud IT services such as Application and Network Layer Security, Disaster Recovery and Backup of the aggregated IOT data;

3.

Large-Scale IOT enabled services such as Home Automation, eHealthCare and Smart Retail which require raw data from IOT endpoints to be aggregated and processed in centralized servers and control decisions from the central servers; for example: turning home automation equipment on or off or setting it to a certain operating metric to be sent back to 1000s (or even millions) of IOT end-points. A Smart Retail use case would involve inventory levels collected from 1000s of product SKUs from retail point of sale locations being aggregated and processed in centralized servers and database engines for subsequent notifications to global supply chain partners, warehouses, etc.

The IOT Themes section discusses a set of use cases which require a coordinated combination of Edge Cloud and Datacenter Cloud capabilities to implement large-scale IOT use cases in enterprise and service provider environments. X-raying into a typical IOT Datacenter Cloud infrastructure, we note a key set of networking layers (see Figure 9):

Page 18

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 9: Key Datacenter Cloud infrastructure layers

Networking SW Tools

Apps & Tools Datacenter Cloud Infrastructure Layer 4/7 Network Intelligence

Layer 2/3 Networking Platforms

Network Automation, Big Data Analytics, Cloud Services Orchestration Tools

Traffic Management, Security, App Acceleration, etc (e.g. F5, Citrix, A10, etc)

Virtual Network Overlays (e.g. VMware, Cisco, Juniper, etc) 10/40GE Layer 3 Leaf/Spine Switching (Cisco, OEM switches w/ Cumulus SW, etc)

Transport

Optical Transport (10/40/100G) (Ciena, Alactel Lucent, Infinera, etc) 5

Source: Deutsche Bank

1.

A 10/40GE Leaf and Spine switching infrastructure designed to be a “Cloud-Scale” Layer 2/3 physical networking underlay for implementing a fully meshed interconnection of application servers running in the datacenter. A Leaf and Spine based Datacenter Cloud can scale to support 10s of 1000s of application servers which is adequate to support a majority of IOT use cases in large enterprise and service provider networks. Cisco’s recently launched 10/40GE Nexus 9k switches and competitive alternatives from HP, Juniper, Brocade and the white box switching vendors (running SW stacks from Cumulus, Big Switch, etc.) are examples of Layer 2/3 network solutions for implementing Datacenter Clouds.

2.

While many IT vendors (Cisco, VMware, HP, Juniper, etc.) support virtual network overlays for running Cloud applications in logical networking clusters (for example: interconnecting a database in server rack A to a Web application in server rack B and to an Analytics application in server rack C), Cisco’s Nexus 9k switches support a virtual network fabric – termed Application Centric Infrastructure (ACI) Fabric – which enables applications running in a Datacenter Cloud to inform the switching and routing layers (via APIs) about their network requirements.

Deutsche Bank Securities Inc.

Page 19

6 May 2014 TMT The Internet of Things

For example: using Cisco’s ACI Fabric, an IOT application running Big Data Analytics in a Datacenter Cloud – based on raw data generated from 10s of 1000s of wireless sensor end-points – could leverage the ACI Fabric to ensure that raw data from the sensor end-points reach the Analytics application servers in a time-sensitive manner. Another example is an IOT application in an Oil and Gas company which could leverage the ACI Fabric to transport Gigabytes of machine data from offshore oil rigs to centralized databases on a besteffort bandwidth available basis since the oil rig data is not timesensitive but is bandwidth-heavy. 3.

Datacenter Clouds are ideal for running Layer 4/7 network intelligence and security features that can be consumed by IOT end-points as Cloud services. Examples of Layer 4/7 feature-functions that can be delivered to IOT end-points as Cloud services are: Application and Network Firewalling; Web Traffic Steering; Per-Session IT Policy and Bandwidth Enforcement, Billing Metrics, etc. Layer 4/7 appliances in a Datacenter Cloud are implemented as either SW modules on vendor-specific HW platforms (e.g. F5’s BIG IP or modular Viprion HW running a wide range of SW Layer 4/7 modules) or as SW modules running on VMs in industry standard x86 servers. While use cases such as security processing (SSL Offload) are likely to require HW optimized vendor platforms, an increasing number of Layer 4/7 feature functions are likely to be implemented as virtualized SW modules running on industry standard servers in our view. Our view is based on our IT conversations noting the operational benefits around automated SW module deployment on x86 servers, ease of managing 100s of virtual SW instances, creating service chains based on a lego-block combination of Layer 4/7 features, etc. We note that Telco Cloud environments – implementing the Network Functions Virtualization approach (Telco NFV; refer to our recent S2N notes on NFV) – are more likely to run Layer 4/7 features in virtualized SW modules for implementing IOT use cases such as e-HealthCare and Home Automation. Fortune 500 enterprise IOT use cases are more likely to use performance-optimized HW platforms for running Layer 4/7 features, given their ruggedized form-factors (e.g. flame, temperature range and electrostatic hazards tolerance, etc) and their relevance in industrial IT environments.

4.

Information Technology (IT) and Operational Technology (OT) Convergence: A centralized Datacenter Cloud in an IOT architecture is ideal for implementing a converged IT and OT infrastructure (see Figure 10). Figure 10 illustrates a typical IOT Datacenter Cloud containing the IT layers, i.e. the physical Layer 2/3 underlay network, virtual overlay networks between racks of servers, Layer 4/7 and security appliances, etc.; and the OT layers, i.e. the industrial control system network and the supervisory network.

Page 20

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 10: Conceptual view of an IOT Datacenter Cloud – illustrating the Information Technology (IT) and Operational Technology (OT) infrastructure elements

Layer 4/7 & Networking SW Features: • Application Acceleration • Policy Enforcement • App + Network Security • Automation • Analytics • Visibility • Orchestration & Service Chaining

Wide Area Networks (Internet, Datacenter Interconnects)

Layer 3 40GE Spine Switch (e.g. Cisco, Juniper…)

Leaf/Spine Cloud Scale Network

Layer 2/3 10GE Leaf Switch (e.g. Cisco, Juniper…)

Server Racks with Virtualized Workloads (e.g. Cisco, HP, Dell…)

Information Technology [IT] Layer Operational Technology [OT] Layer IOT Control Systems Network Programmable Logic Controllers; Actuators….

IOT Supervisory Network SCADA, etc

Source: Deutsche Bank

SCADA – an acronym for Supervisory Control and Data Acquisition – is an industry standard architecture for: 1) acquiring telemetry data from industrial objects such as sensors, etc; 2) converting the sensor data to digital form for delivery over IP networks; and 3) processing sensor data in centralized servers for relaying back control instructions to the sensors. SCADA and other systems management protocols are used in IOT environments for controlling industrial devices such as actuators in smart manufacturing and retail environments, connected transportation, oil and gas installations, home automation systems, etc. Our objective with this simplified overview of Internet of Things Technologies is to highlight the core value proposition of Internet of Things use cases to usher a functional convergence of Information Technologies and Operational Technologies, i.e. IT and OT systems co-existing and working in concert in a Datacenter Cloud in order to implement a variety of IOT use cases. Deutsche Bank Securities Inc.

Page 21

6 May 2014 TMT The Internet of Things

Architecture The current IOT landscape is characterized by the presence of purpose-built applications operating on a network architecture designed (almost exclusively) for the application. As IOT systems progress, they will need to integrate with other systems and this is where the key challenges remain. Currently, these applications operate in silos; they have a distinct architecture that does not lend itself to inter-system communication and operation. A major issue is the presence of numerous protocols, developed to run specific applications, which makes application integration and their respective architectures inadequate and complex. To achieve the scale and level of integration that IOT requires, it will first need a common foundation from which to build sub-architectures – a common set of building blocks for developing system architecture will ensure a high level of interoperability at various levels. Given what is at stake, and every vendor’s partial view to one platform or another, the acceptance of a common architecture is crucial and likely only possible if done from an objective, thirdparty perspective.

A major issue is the presence of numerous protocols, developed to run specific applications, which makes application integration and their respective architectures inadequate and complex.

Currently, various Standards Development Organizations (SDO) are involved in developing reference architecture for IOT. The International Telecommunication Union (ITU) has proposed the ITU-T model and IOT Global Standards Initiative (GSI) to drive standards across all domains of the IOT. oneM2M Seven global SDOs which publish telecom standards came together to form oneM2M in July 2012. The participants were: Association of Radio Industries and Businesses (ARIB); Telecommunication Technology Committee (TTC) of Japan; Alliance for Telecommunications Industry Solutions (ATIS); Telecommunications Industry Association (TIA) of the USA; the China Communications Standards Association (CCSA); the European Telecommunications Standards Institute (ETSI); and the Telecommunications Technology Association (TTA) of Korea. Their overarching objective is to develop specifications to ensure global functionality of systems and minimize fragmentation of M2M and IOT architectures. Recently, oneM2M announced that it will publish its first release of standards in August. The specifications developed by oneM2M will serve as a common structural design for networks and systems to run a broad set of applications. The oneM2M architecture intends to develop a common service layer that facilitates interoperability between sensor/device networks and cloud‐based applications.

Page 22

The specifications developed by oneM2M will serve as a common structural design for networks and systems to run a broad set of applications

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 11: Layers of a generic oneM2M architecture

There will also be a number of communication channels – machine-to-machine, machine-to-human, machineto-application-server – making it almost impracticable to have one, Source: oneM2M

universal solution for all scenarios in which IOT

IOT-A Concurrently to oneM2M, the European Commission established IOT-A, which created the Architecture Reference Model (ARM) and defined the key building blocks of (almost all) IOT systems. These building blocks are referred to as architectural views: Physical view, Context view, Functional view, Information view and Deployment view.

applications are likely to be deployed…

Unfortunately, a single network system for IOT cannot be established using the above views, given the wide range of use cases, differences in physical enddevices and various ways these elements will be connected (Wi-Fi, cellular radio, Bluetooth). There will also be a number of communication channels – machine-to-machine, machine-to-human, machine-to-application-server – making it almost impracticable to have one, universal solution for all scenarios in which IOT applications are likely to be deployed. …Therefore, the architecture

Therefore, the architecture reference model proposed by the IOT-A provides a set of abstractions to develop concrete architecture for a large variety of IOT use cases. IOT-A has suggested the above mentioned architectural views to be implemented in contextual manner. They will serve as the building blocks needed to arrange and configure specific applications. All of this to say that despite the existence of a large number of sub-architectures within the broad IOT framework there will be a common set of architectural layers (building blocks) underlying each framework. Thereby, establishing a common foundation.

Deutsche Bank Securities Inc.

reference model proposed by the IOT-A provides a set of abstractions to develop concrete architecture for a large variety of IOT use cases.

Page 23

6 May 2014 TMT The Internet of Things

Figure 12: Architecture proposed by IOT-A

Source: IOT-a, European Commission

The physical entity view of this is perhaps the most complex of all the layers of IOT architecture. It consists of broad range of physical objects from license plates to orchids, jet engines, thermostats and anything that could be relevant from user or application perspective. Therefore, different IOT systems will need to have specific provisions for physical objects they will address. It is also important to give adequate consideration to physical objects connected to IOT when planning system architecture, as different physical objects and the applications designed for them can greatly influence the system architecture. Physical objects differ from each other in terms of their properties, dimensions, functionality and purpose (see the license plate and orchid mentioned above). Consequently, different physical objects require different types of sensors, actuators and other monitoring tools that influence the system architecture in a different manner.

Different physical objects and the applications designed for them can greatly influence the system architecture

The way physical objects interact with sensors and actuators and what type of information is gathered by them is also an important consideration by standardization bodies. Whether sensors are attached to the physical objects, remotely monitoring them, in close proximity, or inside the physical entity? This further influences the approach to system architecture. Reference architecture for IOT extends to Wireless Sensor Networks (WSNs), which were originally developed for military and heavy industrial applications (not necessarily with the intention to bring them to the consumer domain). Early on, the United States Military developed the Sound Surveillance System to identify and locate submarines from the Soviet Union using submerged acoustic sensors. The architecture used for the Sound Surveillance System still Page 24

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

resembles the ones used in modern WSN deployments to monitor volcanic activity. Later, the United States Defense Advanced Research Projects Agency’s (DARPA) investment in Distributed Sensor Network (DSN) program in 1980s, brought academic research to the field of wireless sensor networks via participation of MIT and Carnegie Mellon University. Following this research, the government and other institutions started using WSNs for civilian use. Some of the early civilian applications of WSNs were for the prevention of natural disasters, air quality monitoring, building structure monitoring and forest fire detection. Commercial deployments were focused on industrial uses like factory automation and efficient power distribution. Despite the successful implementation of wireless sensor networks in commercial and public places, the use of proprietary standards made it difficult to extend their reach beyond a narrow scope of applications. As with the early computing industry, WSNs were focused on performance and functionality rather than power consumption and scalability. In order to accelerate the pace of growth there was need to standardize the protocols, as well as reduce the size and lower the cost of the sensors. As it stood, early WSNs never overcame these challenges and were limited to mostly government use. Figure 13: WSNs market traction as sensor costs decrease

Source: Silicon Labs

A vast number IOT applications will rely on wireless sensor networks for their functionality. Many deployments will likely be in remote areas with no readily available power source, requiring sensor nodes to go without a charge for days, weeks or even months. Some deployments will have the added challenge whereby they are in areas with limited network connectivity. As IOT pulls WSNs into the consumer domain, cost constraints will be an additional hurdle, requiring the new generation of sensors to not only be resource efficient but cost effective.

Deutsche Bank Securities Inc.

Page 25

6 May 2014 TMT The Internet of Things

Wireless sensor networks will make up a majority of the elements used in IOT, enabling physical objects to be connected to applications and cloud services. Data gathered through these sensors will be analyzed in various ways using cloud computing and Big Data approaches. Like many other components of IOT, wireless sensor networks are evolving at a rapid rate. Though contemporary WSNs capable of being deployed in a wide range of scenarios, their present-day implementations are limited and characterized by proprietary development environments and interfaces.

Wireless sensor networks will make up a majority of the elements used in IOT, enabling physical objects to be connected to applications and cloud services

Much like the standardization of the architectural approach to network systems for IOT, wireless sensor networks also need a holistic architecture model, as the heterogeneous nature of present-day deployments can be a barrier to ubiquitous WSNs. Ideally, this will ensure the seamless transfer of data from sensor nodes to applications and the ability of WSNs to integrate with each other and also larger IT systems and the internet. There are and will continue to be challenges with sensor nodes, especially given the number. For one, they are typically resource-constrained in wireless networks. Moreover, they have low processing capabilities, less memory availability and need to operate with a low energy footprint. This makes it difficult to deploy WSNs and integrate them with IP based services.

There are and will continue to

The majority of WSN deployments are currently proprietary. Designed for specific applications and services, they cannot be replicated when other applications and services are considered. As a result many wireless sensor networks are isolated deployments. Eventually though, this will lead to redundant WSN deployments. For example, if a wireless sensor network was deployed to measure temperature and water level in a field and is later required to monitor sunlight and humidity, the existing WSN cannot be adapted to measure additional information, leading to the development and deployment of an additional WSN.

in wireless networks.

be challenges with sensor nodes, especially given the number. For one, they are typically resource-constrained

Tracy and Sreenan architecture An additional architecture was proposed at IEEE in 2013 by David Tracey and Cormac Sreenan of University College Cork, Cork, Ireland. According to Tracey and Sreenan, the reference architecture of the wireless sensor networks is likely to have multiple layers for different nodes that perform different functions. For example, a node might be engaged in measuring humidity in the environment, while the other one would be engaged only in forwarding the date collected by other nodes and not measuring any physical phenomena. In this case, the forwarder will not have a local instrumentation layer.

According to Tracey and Cormac, reference architecture of wireless sensor networks is likely to have multiple layers for different nodes that perform different functions

The prerequisites of this proposed architecture to develop and deploy highly integrated WSNs is to have an architecture that is agnostic to sensor node functionality and does not require high power consumption. The architecture should also be able to manage small, static networks and allow the system to adapt as the network grows and changes, while at the same time providing a reasonably consistent means to retrieve the information given by the independent and changing sensors. In addition to being agnostic to sensor node functionality and allowing the consistent exchange of sensor information, the architecture should to be able to provide support for integration with other systems for analysis and modeling of sensor data.

Page 26

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 14 illustrates the layers in the architecture for nodes of different capability with their different roles. For example, a node that only fulfills the role of forwarding information would not have an instrumentation layer, but has an object space to store data from remote peers. Figure 14: Tracy and Sreenan architecture

Source: Tracey and Cormac, University of College Cork via IEEE

This architecture uses a simple protocol based on Peer to Peer (P2P) concepts able to run on nodes with limited resources. It is a holistic approach which considers the entirety of the data flow between the actual sensor and the service(s) it is delivering, supported by lower layers, rather than each layer specifying its own behavior in isolation. As a result, it is an orchestrated effort rather than a coalition of random elements.

It is holistic approach which considers the entirety of the data flow between the actual sensor and the service(s) it is delivering, supported by lower layers, rather than each layer specifying its own behavior in isolation

Deutsche Bank Securities Inc.

Page 27

6 May 2014 TMT The Internet of Things

IOT Protocols Protocols are another place where the shifting tides of OEMs opinions will likely play a key role. Currently, there are a number of protocols developed and in development, with a few getting more traction than others. Historically, the standardization process for protocols in other industries has been long and riddled with dead-ends. The closest proxy to the Internet of Things, and likely one that would incorporate many elements, would be the wireless industry itself. When wireless emerged as something that could fork into separate technologies, largely one standard body won over another because the technology simply worked better (WCDMA) or was significantly cheaper (GSM); in some cases, however, even the best technology may not win out, as negative marketing campaigns can undermine a good idea. And given there were a few standards bodies competing for their own protocol standard to win out, this did happen.

Historically, the standardization process for protocols in other industries has been long and riddled with dead-ends.

To get a sense of how numerous protocols may affect the implementation of IOT, imagine a room filled with people, each speaking a different language, and none understanding anyone else. It would be a room full of noise with little if any communication. A scenario similar to the one could play out with connected devices, as most function on proprietary standards, thereby limiting their interaction with other devices. For example, different applications were created to allow an Android device to communicate with a Samsung smart TV, versus having that same TV controlled with a remote control operating on the Zigbee interface. It is likely that those two devices, the remote and the phone, have little interaction with each other – not that that would be imperative in this case; however, it could be with other use cases. Figure 15: Managing disparate devices/services and transport layers What They Say

What They Hear

God dag!

Hello Son!

God Dag! Hello Son! God Dag!

Kurt Hyvää päivää!

Transport layer

Hello Son!

Paul

Dag! HelloGod Son! God dag!

Hyvää päivää! Hello Son!

Jorma Hyvää päivää ! Hello Son! Hello Son!

Sten

Hyvää päivää!

Hyvää päivää!

Pekka

Irwin Devices/Services

Source: Deutsche Bank

Page 28

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

This is where industry standards, protocols and hopefully the resulting interoperability play a critical role. Typically, in order to find a commonplace among disparate interfaces, protocols are put in place. By definition, protocols are a set of predefined rules used by devices in a network to communicate with each other. A typical communications network has multiple layers of protocols. Each required for a different set of devices and tasks, and each becoming an essential element of the network, as they govern how data transmission and communication between devices on a network take place.

A typical communications network has multiple layers of protocols. Each required for a different set of devices and tasks, and each becoming an essential element of the network, as they govern how

For the Internet of Everything to move beyond a hype cycle, all of its components have to work in a platform agnostic, OS agnostic and transport layer agnostic manner. There have to be protocols that devices conform to, no matter which OS they operate on and which company manufactures them. Just the way knowledge of a common language would make it possible for all the people to communicate in our previous example, no matter which country or region they come from.

data transmission and communication between devices on a network take place.

No clear protocol in place If a protocol can be defined and generally accepted early in the development of a platform, then the chances of the platform’s success are significantly improved. The Internet is one example, as the standardization of Internet protocols were the main stay of growth in web-based applications. The Internet of Everything, however, is yet to find a standardized set of protocols. Various industry groups and associations are engaged in developing protocols that will serve as framework for transmitting data by devices and sensors in the Internet of Everything to each other and the network. Given the magnitude of connections applications emanating from IOT, it is nearly impossible for one company to develop standards and protocols. Therefore, various industry alliances have emerged to address this need. Driven by the industry effort, there are a number of acronyms of protocols being touted to become the standard for IOT. We list most of the existing protocols and place what we believe to be their chance of success. We measure this chance of success based on being light weight (in terms of code footprint), low power consumption, open and suitable to a multi-platform environment. „

AllJoyn

„

CoAP

„

BACnet

„

802.15.4e

„

6LoWPAN

„

AMQP

„

XMPP

„

DDS

„

RPL

„

RESTful HTTP

Deutsche Bank Securities Inc.

We measure the chance of success based on being light weight (in terms of code footprint), low power consumption, open and suitable to a multi-platform environment.

Page 29

6 May 2014 TMT The Internet of Things

„

Dash7

„

DTLS

„

UDP

„

ZigBee Pro

„

RFID

„

Bluetooth

AllJoyn Originally developed by Qualcomm, Alljoyn is, as the name suggests, an open source project intended to provide an eventual universal framework to enable interoperability among devices and software applications. Qualcomm has handed the reins to Linux, and the open source project is now overseen by the AllSeen Alliance, which is a non-profit consortium. The move to a consortium oversight has encouraged participation by others, and the list of both partners and developers has increased significantly since it made this move outside of Qualcomm’s umbrella. As shown in Figure 16, the list is lengthy but noticeably does not include prominent semiconductor names like Intel and ARM. Intel is currently building a protocol around Quark and their Atom 3800, parallel to the way it handled the wireless standards with its promotion of WiMAX. However, the AllSeen list includes Cisco, an encouraging addition given Cisco’s own IOT initiative and backing of MQTT (see below). Ultimately, the goal of the Allseen Alliance is to push participants to devote engineering resources to develop code for the open source framework and eventually enable disparate devices and services to discover, connect and interact, regardless of transport layer. A key directive for AllJoyn services and applications is that they are created with a source code which is transportlayer agnostic, be it powerline, wireless or Ethernet, and leverage this in a manner indifferent to manufacturer or internet connection.

Figure 16: Members of AllSeen Alliance Haier LG

Harman

Panasonic

HTC

Qualcomm

Control Networks

Sharp

Imagination

Silicon Image

Kii

Technicolor

LIFX

TP-Link

Liteon

2lemetry

Moxtreme

Affinegy

Musaic

ATT

Muzzley

Cambridge Audie Beechwoods Winner Micro CA Engineering Canary Cisco

AllJoyn was developed as a mesh networking service which offers automatic discovery and communication for a number of different devices, agnostic of operating system. Even though the Allseen Alliance is part of the Linux Foundation, the goal of AllJoyn is to be cross-platform with support for Android, OS/X, Windows variants, gaming engines and other thin clients.1 While it does not support Real-time operating systems – those that process data without buffering delays – some of those operating systems mentioned above are likely to eventually be more controller-like, with low-end embedded duty as enhancements are made with Internet connectivity and overall intelligence. Eventually, the protocol will eliminate the need for a hub, or an Internet connection.

Gowe

D-Link

Patavina Technologies Sears Sproutling The Sprosty Network Tuxera Twobulls Vestel

Double Twist

Weaved

Fon

Wilocity

Source: AllJoyn

The reality of an automated home environment driven by AllJoyn devices may not be too far away. We were able to see what this environment, enabled by AllJoyn devices, may look like in the near future during Mobile World Congress. Qualcomm had part of its booth set up demoing the possibilities. Given the ample YouTube video of the booth tour, we will not recount every detail here. Suffice to say that AllJoyn enabled everything from door locks to wine refrigerators and teddy bears to both send and receive event triggers and

1

http://linuxgizmos.com/open-source-iot-inititiative-taps-qualcomm-alljoyn-framework/

Page 30

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

notifications in a peer-to-peer manner, across a number of transport layers and across different device manufacturers. We believe AllJoyn to be the largest, cross-platform protocol effort existing today, given the number of players involved and their respective clout, as well as the oversight from the Linux foundation. Soon, we expect a number of connected, in-home devices designed with AllJoyn, beginning with LG’s new smart TV announced late last year. However, despite the backing there are still other protocols, either industry specific, device specific or transport layer specific and some agnostic to all of these, and while AllJoyn has done a decent job of generating support for their consumer-oriented devices, there are still other protocols which have gained support around other use cases, or will simply work better for specific applications. We walk through a few of the important ones below, as well as provide a general description of some of the others. MQTT – Message Queuing Telemetry Transport Originally developed by IBM and Arcom (now Eurotech) in 1999, MQTT is a lightweight network connectivity protocol used to publish/subscribe messaging between devices. The term ‘Telemetry’ in the definition refers to automatic measurement and collection of data, often from inaccessible or remote locations. Data collected is transmitted to the receiving equipment for monitoring and further processing. MQTT has a small code footprint, thus making it appropriate for use in remote sensors with limited processing capability and memory availability. Being lightweight also helps in situations where bandwidth is scarce and latency is not imperative. Put together, these factors make it possible for a large number of clients (or things in Internet of Things) to be connected to a single server. MQTT also has something called the Extended Reach component, which facilitates connections between the messaging backbone and devices at the edge of the network. It allows a wide variety of devices from mobile handsets to trains, refrigerators, health care devices, smart meters, door locks and cars to transmit data using sensors for processing and analysis. Embedded sensor instrumentation facilitates real-time analysis of data generated by remote sensors.

MQTT has a small code footprint, thus making it appropriate for use in remote sensors with limited processing capability and memory availability

Having processed data, the central processing unit then communicates back by sending instructions to devices to act in the optimal manner based on certain situations.

Deutsche Bank Securities Inc.

Page 31

6 May 2014 TMT The Internet of Things

Figure 17: MQTT machine-to-machine communication

Source: IBM

MQTT protocol was presented to OASIS for standardization in February 2013, and based on initial working specifications of the protocol it is likely to be completed by March 2014. OASIS is a non-profit consortium that develops open standards for communication technology. Members of the Technical Committee at OASIS for standardization of MQTT include Cisco, Eclipse Foundation, Eurotech, IBM, Machine-To-Machine Intelligence (M2Mi), Red Hat and others. CoAP – Constrained Application Protocol CoAP is an application layer software protocol for device-to-device communication. It is meant to be used by small devices like switches and sensors. These devices are generally constrained by resources, typically using only 10 Kb of RAM and 100 Kb of ROM, 8-bits of microcontrollers and low quality networks with high error rate. Since small devices are constrained by resources like processing power, storage space, power and generally a poor network connection, CoAP is designed to be simple and light weight.

CoAP is meant to be used by small devices like switches and sensors.

CoAP enables integration of devices/sensors with the web by providing easy interface with HTTP and allows devices to have request/response interaction. CoAP works with devices that support either UDP or UDP analogue (UDP – User Datagram Protocol is used to transport compact messages). CoAP has two sub-layers: i) a message layer that works with UDP to ensure duplication detection and reliable delivery of messages; and ii) a request/response layer which enables interaction through GET, PUT, POST and DELETE functions.

Page 32

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 18: Abstract layering of CoAP Application Sub-Layer Requests/Responses

CoAP Messages Sub-Layer UDP

Source: IETF

Figure 19: Architecture of a CoAP-based wireless sensor network

Source: mdpi.com

CoAP is currently being used as the enabling technology for electric utility AMI (Advanced Metering Infrastructure) and DI (Distributed Intelligence) applications within Cisco’s Field Area Network. Features of CoAP: „

auto discovery of resources

„

enables native push notifications and simple subscription for a resource

„

minimizes the complexity of mapping with HTTP

„

lowers the header overhead and parsing complexity

„

enables content-type support and URI (uniform resource identifier is a string of characters used to identify a name of a web resource)

„

enables the device to act as both client and server.

Deutsche Bank Securities Inc.

Page 33

6 May 2014 TMT The Internet of Things

6LowPAN Low power network and low power devices – 6LowPAN addresses the need of low power devices in low power network to enable IPv6 communication over short range wireless personal area network. 6LowPAN is an acronym for IPV6 over Low Power Personal Area Networks. It enables sensors that have constrained processing capability and restricted energy resources to communicate over IPv6 networks. Published in 2007 by the IETF, 6LoWPAN optimizes communication by compressing message size over low-power radio technologies such as IEEE 802.15.4. For instance, 6LowPAN compresses 60 bytes of headers to 7 bytes, thereby facilitating sensor and device communication over highly constrained networks.

6LoWPAN optimizes communication by compressing message size over low-power radio technologies such as IEEE 802.15.4.

Along with MQTT, COAP and Alljoyn, 6LowPAN is also receiving industry support as a relevant protocol for IOT based communication. In October last year, IBM and Libelium launched 6LowPAN development platform for Internet of Things to provide Internet connectivity to sensors and devices via IPv6 protocol. The goal of 6LoWPAN is to include low power devices in Internet of Things, and the implementation of IPv6 over IEEE 802.15.4 standard helps in achieve that goal. IPv6, the latest version of Internet protocol, is a successor to IPv4. In the realm of IOT, IP addresses have gained importance as they make it easier for these devices to communicate with the Internet and each other. In early 2011, the last blocks of IPv4 internet addresses were allocated, marking the shift in the uptake of IPv6 which not only offers additional addresses but also brings more comprehensive security features. 6LowPAN allows sensors to be easily established by plugging into standard IP sockets. Furthermore, 6LowPAN networks are suitable for mesh networking, facilitating the deployment of a large number of sensors with relatively fewer gateways, thereby simplifying the architecture and reducing capital expenditures.

6LowPAN networks are

Devices built on IEEE 802.15.4 are designed to have small form factor, low energy footprint, low cost and flexible installation (think wearable devices). While IEEE 802.15.4 standards focus on code-size optimization, IPv6 aims at achieving high transmission speeds; 6LowPAN helps bridge the difference in focus of two important standards.

gateways, thereby simplifying

Page 34

suitable for mesh networking, facilitating the deployment of a large number of sensors with relatively fewer the architecture and reducing capital expenditures.

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 20: 6LowPAN enables IPv6 over low power WSNs

Source: www.eis.ernet.in

Outside of the protocols mentioned above, there are a number of other protocols. Most are application specific, and many could be eventually integrated into the larger, holistic protocols listed above. As such, we provide a short description of each below: BACnet: BACnet is a protocol developed by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE). It is used to build automation and control networks and is an ISO global standard that is used in more than 30 countries. IEEE 802.15.4: A standard maintained by the IEEE 802.15 working group for low-rate wireless personal area networks to specify physical layer and media access control. AMQP: An open source protocol used to enable devices for message oriented communication between them. Data Distributed Service (DDS): An open source standard for middleware in real-time and embedded systems. It facilitates exchange of data in a scalable, dependable and interoperable manner between publisher and subscriber. RPL: It is a routing protocol specially developed for low power devices and helps in best-path communication, with an idea towards resource constrained devices. RESTful HTTP: Representational state transfer (REST), which runs over HTTP, is a simple architecture which is used to read XML file of a web page. A XML file of a particular page includes the desired content and describes it. DASH7: Developed by DASH7 Alliance, DASH7 is a long range and low power open source standard for wireless sensor networking. Datagram Transport Layer Security (DTLS): DTLS is a protocol which provides security and privacy to datagram protocol. As basic datagram protocol delivery

Deutsche Bank Securities Inc.

Page 35

6 May 2014 TMT The Internet of Things

of data packets is not guaranteed, DTLS is commonly used in delay sensitive and security concerned applications. User Datagram Protocol (UDP): Developed by David P. Reed, UDP is a transport layer protocol. It is used to transport datagrams over Internet protocol. ZigBee Pro: ZigBee is a specialized protocol for low power devices. It enables low power devices to achieve high level of communication to form personal area networks and local area networks, typically with low-cost devices like remote controls, for applications such as home automation. Silicon Labs has an extensive history of creating low-power processors for IOT elements, many of which use the Zigbee Pro protocol. The Zigbee Pro protocol is a feature set which offers connectivity to Zigbee enabled devices but is also easier to use and able to support a larger number of devices in one area, than its predecessor, Zigbee. Radio Frequency identification (RFiD): RFiD is way of transferring data through radio frequency electromagnetic fields emitted by small devices with an antenna and a small chip. Currently, it is used to automatically track and identify tags containing information about physical goods needing to be tracked. Impinj, a private company with over $100 million in funding, has focused on a technology called ultra high frequency (UHF) RFiD, which has longer range, higher speed and lower cost than traditional RFiD. Bluetooth: Developed by Ericsson, Bluetooth is a personal area network protocol used to exchange data wirelessly over short distance between devices. It has gained wide adoption given its minimal energy requirements and high fidelity over short ranges.

Page 36

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Security Security is a key hurdle for IOT to gain widespread adoption. Without a coherent, holistic and integrated security mechanism in place, the risks of deploying IOT, given the potential for malicious attacks, could easily outweigh the benefits. Given the lack of a common protocol, the task to secure disparate entities remains daunting. To understand, think about the number of times the press discusses some weakness or a large attack on Internet security – Heartbleed, Target credit card breech, daily DDOS attacks – and this is limited to a framework which has a common protocol. Now, imagine dealing with a network of resource-constrained devices, all of which will always remain on, connected, distant and likely autonomous. Given the business models concocted just to undermine the Internet, it is not difficult to extrapolate the risks inherent to IOT and the undertaking to secure such a platform. Given the risks, security must be implemented into the foundation with a holistic view for all elements and process.

Without a coherent, holistic and integrated security mechanism in place, the risks of deploying IOT, given the potential for malicious attacks, could easily outweigh the benefits.

Network security Given the variety of protocols in play and the resulting heterogeneity of the overall network, the resulting diversity will severely affect any security mechanism’s ability to protect the overall platform. Securing any communication channel, from remote devices to controlling devices for example, is difficult considering how resource constrained these remote devices will be. For example, the addition of a simple key management systems may bog down the system, a system which needs to be light and flexible. Nevertheless, some still argue that cryptography, if built into the network infrastructure at the foundation level, could be a resolution2. Cryptography is the transformation of data into an indeterminable format and requires a key to unlock the message. The argument is that cryptography is like mortar, associating unrelated items in a resource-constrained and dynamic environment, and therefore the keys need to be changed in order to protect the network over the long term. It is likely that regular Internet security protocols can be used further up the communication stack, the risk being that the differences between Internet protocols and IOT protocols could lead to vulnerabilities in a system meant to be comprehensive. Data protection It is a challenge to conceptualize the amount of data generated by one person in an idealized IOT world, from payments and in-store beacons to home automation and personal health monitoring, but taking into account the current use cases it is easy to see how a Big Brother-type environment can be developed, as it already has in some geographies. We will leave the philosophical discussion on what should and should not be accessed by the government for others to address. However, the overriding concern relating to the Internet of Things and security is that given the knowledge that someone can and likely will access this data, to what degree are consumers willing to leverage the conveniences of IOT? We believe that even the most trusted sources for security will not likely be fully trusted by the majority of users and

2

We believe that even the most trusted sources for security will not likely be fully trusted by the majority of users and will likely depend on the respective application’s personal reach as to what degree it has the potential to become ubiquitous.

IEEE Computer, vol.44, no.9, pp. 51-58, September of 2011

Deutsche Bank Securities Inc.

Page 37

6 May 2014 TMT The Internet of Things

will likely depend on the respective application’s personal reach as to what degree it has the potential to become ubiquitous. Yet, IOT will place an incredible amount of personalized data in the general environment. There are some possible solutions that could assuage fears: First, the amount of data accessed, or exactly which data is accessed, could be curtailed or refined by the user (think of it as an extension of the alerts one already receives when specific applications want location-based information). Second, the information could be stored by trusted sources. Admittedly, this would be challenging, but could be helped if the process is transparent. In short, data protection is the key to user adoption and could be helped by the steps mentioned above. However, the adoption rate of IOT will be a series of fits and starts, bound by the number and degree of data breaches, which will happen. Managing identification Devices will continue to grow and each of them needs to have the ability to be identified. Inanimate objects such as cars or warehouses have the same issue; there are a number of sensors associated with each object – static or dynamic (e.g. a sports arena which changes into a rescue shelter during a catastrophic event, and hence the identification of this space needs to change dynamically). In short, a number of IDs would be associated with each device or a group of devices, and proving these IDs through authentication and authorization, and within an architecture, which will in many cases be a distributed architecture, would be a challenge. Hence, for any embedded security element, the number of elements, their relationship and the potential to have them changed will need to be considered.

The adoption rate of IOT will be a series of fits and starts, bound by the number and degree of data breaches, which will happen.

Mechanisms in development There are a number of standards being developed by industry bodies, researchers and governments, and these are listed below. All are decent first steps but without a true ramp up in devices and sensors it will be a process of incremental change to address aspects which are currently unknown. „

Ucode: identification number system which identifies objects and places uniquely (www.uidcenter.org)

„

GS1 keys: ID system used for items, services locations, containers (www.gs1.org)

„

IEC 62591: protocol for wireless sensors with encryption, authentication management and key management

„

ISO/IEC 14443: overall architecture for contactless proximity cards which provide information flow protection3

As we discussed earlier, cryptography is an important element of security, but an element which will be embedded inside the protocols on which the devices run. These are being developed from the ground up, as it is a challenge to implement Internet protocols like IPsec to provide a holistic, transport layer security blanket. This is where the aforementioned list of security mechanisms will target.

3

IEEE Computer, vol. 44, no. 9, pp 51- 58, September 2011

Page 38

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

In total, security will need to combine the challenges associated with resourceconstrained devices, varying transport layers and likely different protocols. Cryptography mechanisms will help, but challenges around ID management, user privacy, and the lack of a single, trusted architecture are formidable hurdles that would have to be cleared for IOT to work in a seamless manner. Additionally, governments will need to provide a balance of oversight and protection of its citizens in the most transparent manner before IOT is likely to reach its potential.

Deutsche Bank Securities Inc.

In total, security will need to combine the challenges associated with resourceconstrained devices, varying transport layers and likely different protocols.

Page 39

6 May 2014 TMT The Internet of Things

APIs Without Application Programming Interface (API), the development of applications may not reach a critical mass. API is a communication interface, which specifies how software and other components should interact with each other using a set of instructions and standards for a given platform. Almost every software and web-based company releases a set of APIs to developers. These APIs enable developers to create applications relatively easily, benefitting the companies and developers. APIs are also helpful when developers are programming the user interface component of applications.

APIs specify how software and other components should interact with each other using a set of instructions and standards for a given platform.

APIs have become a critical element of web, enterprise, mobile, Big Data and IOT ecosystems. The use of APIs extends to multiple environments and enables a broad spectrum of functionality, from basic functions such as copy and paste of text to more complex tasks such as online transfer of funds from a bank account. An example that illustrates the role of APIs is Google and its use of maps and videos. Google provides APIs to developers in order to incorporate its webbased services such as maps on a developer’s website. Using an API for Google maps, developers can embed map functionality to their website and seamlessly provide some of the mapping solutions without having to redirect users to Google’s website. Likewise APIs for YouTube allow developers to embed YouTube videos to their respective websites or applications. In a simple use case scenario we demonstrate the use of Google’s maps application by Verizon. For users looking for directions to Verizon’s stores, the company website leverages Google’s map data to provide instructions. In the store locator section of Verizon’s website, users can click on ‘Get Directions’ link and be directed to another Verizon webpage with embedded Google map and directions to the store.

Page 40

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 21: Google maps via Verizon’s website

Source: Verizon, Google

Figure 22: Directions to store via Google maps

Source: Verizon, Google

Another example is Amazon that provides APIs to web developers, allowing them to access product inventory and updated pricing details. A developer can use Amazon's Product Advertising API to list Amazon products with updated prices on his website and possibly monetize the sales opportunity.

Deutsche Bank Securities Inc.

Page 41

6 May 2014 TMT The Internet of Things

APIs run in the background, not necessarily visible to end users. Users interact with the application, which in turn interacts with other applications using APIs. In addition to governing the way applications interact with each other, APIs also govern the information which applications can exchange with each other. A developer can extend the reach of his application and incorporate functionalities such as gathering or sharing feeds from social media platforms, carrying out analytics, building widgets such as Yahoo weather or including messaging capability to their app. In doing so, APIs make it possible to create application mashups which are simply analogous to musical mashups where two or more unrelated songs are blended together in a creative mix. Likewise, in a mashup of web-based applications, functionality of two or more applications are brought together using APIs to create a new application. In this fashion, APIs can be used to extend functionality of IOT applications (see Figure 23).

In addition to governing the way applications interact with each other, APIs also govern the information which applications can exchange with each other.

Trendsmap.com combines Twitter and Google maps to inform users about popular topics in the locale they are located. When we looked up New York on Trendsmap.com, it was easy to see ‘tgif’ and @ladygaga among the most popular topics on twitter. This is a basic example of a mashup, but it shows that creative and useful applications can be developed without requiring developers to write lengthy codes. When Google created Google maps it might not have imagined that it would be used to lookup popular tweets or multiplayer games. We believe that mashups for Internet of Things will open the door for creating new and useful services without the need for them to be thought up at the outset. Figure 23: Mashups make it relatively easy to combine applications

Source: Trendsmap

Page 42

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

In a nutshell, APIs have become a lynchpin of sorts in the process of application development in the Web 2.0 era. As they have evolved, the use of APIs extended from web-based applications to IOT applications. Web APIs are one of the key enabling technologies of IOT, largely given that IOT applications are eventually connected to the web in some way. In addition to incorporating new features, APIs can enable sensor nodes to perform various functions such as data collection, storage, transmission and data processing. They also help configure sensor nodes to log data on various parameters such as temperature, location, pressure, humidity and other measured elements, and then transmit that data to web services. However, simply gathering data and transmitting it to the backend is not necessarily the end goal. Data needs to be contextually filtered. Sensors should not simply be a data dump as it will only overwhelm the network, like processing redundant data from billions of sensors, thousands of times per day. To put this amount of data into perspective, it is estimated that sensors in a jet engine can generate up to twenty terabytes of data in an hour. Multiply this amount of data for a sixteen-hour transatlantic flight from Dubai to San Francisco and we get a sea of data sent to applications and analysts for just one flight. Most of this data is redundant and can be filtered without hampering the quality of analytics. Sensors, therefore, are expected to have the intelligence to filter data in contextual manner and transmit only the relevant instances of information.

Data needs to be contextually filtered

it is estimated that sensors in a jet engine can generate up to twenty terabytes of data in an hour

In certain cases sensors may be required to process the gathered data locally, while also sending it across in various formats for servers and other web services to process. Additionally, there could be issues related to security and resilience of wireless sensor networks (WSN). Frequently, a solution to these issues will likely rest in API management. In a broader sense, APIs bring secure connectivity and intelligence to otherwise isolated and stranded sensors by facilitating secure data connections between applications and sensor nodes. Using Representational State Transfer (REST) architecture that works through authentication and policy compliant authorization, APIs facilitate large scale integration of resources and simplify application programming. With this, applications connected to sensors could be hosted in cloud, data centers or in any other hosted environment accessible via API. REST is a simple, stateless architecture which draws from HTTP instead of competing with the protocol. In present-day scenario, REST is being extensively used by social networking websites, automated business processes, mobile applications and mashup tools. REST APIs facilitate interoperability among disparate systems (a useful feature as IOT is characterized with presence of disparate systems) and reuse a number of software components. REST also makes it possible for a large number of IOT components to interact with each other, as in the case of its application over the internet.

REST is a simple, stateless architecture which draws from HTTP instead of competing with the protocol.

In the context of IOT, APIs function as a thread that stitches various protocols together to allow applications to connect to devices using different protocols. In current deployment scenarios, a vast majority of applications are able to connect to only those devices that are protocol compliant. APIs such as Smart

Deutsche Bank Securities Inc.

Page 43

6 May 2014 TMT The Internet of Things

Object API make it possible to connect different devices running on different protocols to an application. Figure 25: Enabling interoperability via APIs

Figure 24: Sensor middleware, gateway and API management

Source: Michael Koster via datamodels/blogspot

This requires APIs to become an integral part of wireless sensor networks (WSNs). While most of the new deployments of WSNs have a built-in layer for API management and sensor middleware, legacy WSNs were not designed for this. To bridge the gap between standalone WSNs and the network of sensors with IOT capability, designers need API management tools, gateways and sensor middleware. They help identify contextual information, facilitate interoperability, ensure security and overcoming network disruptions. The API layer is essential to facilitate interaction of sensors with applications, cloud and data centers. In the case where sensor networks consist of nodes with different hardware capabilities and functionalities, middleware can help provide a holistic view of the network. It also provides a greater level of security and encryption in situations where sensors are deployed to collect sensitive information. IOT gateways operate at a level above sensor middleware and provide an interface for applications to interact with data from other applications, devices and APIs. The ability to source data from other related applications and devices allow native applications to have a better contextual view and therefore respond in more efficient manner. Given the relevance of APIs and their management in implementation and growth of IOT, the API space is growing quickly. In terms of how this is playing out monetarily, last year Intel acquired Mashery for an estimated US$180m, CA Tech acquired Layer 7 technologies for an undisclosed amount, and MuleSoft acquired ProgrammableWeb. API management is a competitive market with a number of players and there is still a long runway before a true leader emerges. Figure 26 illustrates the Gartner quadrant as it relates to APIs. While we do not believe that the Gartner quadrant is a definitive guide to leaders, it is a decent starting place.

Source: Intel

Page 44

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 26: Gartner Magic Quadrant

Source: Gartner

Middleware The current infrastructure environment does not fully support what is required to leverage the Internet of Things, in an ideal sense. Middleware can play a role here. It will likely be either specialized devices or software, both of which can enable the Internet of Things by leveraging infrastructure and devices following different protocols. Middleware as both software and devices Middleware devices are mostly based on the gateway concept, a more sophisticated network enabling tool than a switch or router. They are designed to address issues related to interoperability, discovery, scalability, communication and security of the devices and data. Leading middleware vendors include Oracle, IBM and Software AG. IBM provides complete middleware solutions to enterprises. Oracle has middleware range by the name of Oracle Fusion. Software AG provides middleware specific to real-time data management and analytics solutions as it relates to the Internet of Things network.

Middleware devices are designed to address issues related to interoperability, discovery, scalability, communication and security of the devices and data

Middleware Software should control and channel the communication in the network efficiently. Moreover, it should manage tasks like data offload, maintenance, error check and network security. Ideally, middleware software should complement middleware devices as well. To date, the software mostly comes in the form of different APIs, which perform various desired functions, mostly in a piece-meal manner. At its core, middleware addresses a number of issues:

Deutsche Bank Securities Inc.

Page 45

6 May 2014 TMT The Internet of Things

Interoperability: In the current environment, devices use disparate protocols for communication and middleware can bridge this difference. Interface: A common interface should be used, or at least a minimal number of interfaces should be used, so that users do need to educate themselves on basic functionality when controlling new elements. Scalability: Given the number of devices attempting to connect to the network, middleware must efficiently control traffic without compromising scalability. Data: Data will need to be offloaded to the cloud after validation and structuring. Given the constraints with sensor nodes, middleware can help with additional functionality. Security: A powerful middleware could help provide a security solution, or at least limit the damage if one sensor is breached.

Page 46

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

IOT as a Platform The Internet of Things will be a complex structure with various interconnected parts, put together by a wide array of industry participants including system integrators, telecommunications operators, gateway manufacturers, semiconductor manufacturers, software vendors, network equipment providers and sensor manufacturers. These players develop applications and software, manufacture equipment and sensors to collect and send data, analyze it, provide insights and send contextual instructions from remote locations for efficient functioning of end devices and much more. Ideally, any IOT device that labels itself as a platform should work as glue to bring together disparate pieces and fit them together in an efficient and context-aware manner. The platform for the Internet of Things will be more than just a development, management and support environment. In our view, there will be two primary approaches to IOT platform.

Ideally, the IOT platform should work as glue to bring together different pieces and fit them in efficient and context-aware manner

First is the software level approach, wherein the platform will facilitate application development, host connected devices, ensure secure connections and transmission of data, and carry out analytics. The other approach is a hardware-centric approach. This approach will provide a platform to facilitate development of hardware for IOT and will quicken the process of bringing connected devices to market. The software based approach to IOT could make it easier for developers to build applications and services for connected devices, monitor and manage devices, simplify connectivity, collect and analyze data, provide APIs and integrate IOT applications to business systems. Over the past several years, a host of platforms have emerged for the development of M2M and Internet of Things applications. The integrated approach adopted by some of the new entrants could disrupt the traditional M2M platform types, as integrated platforms lead to faster development of applications compared to traditional platforms. In addition, integrated platforms host connected devices, enable development of scalable and portable applications and manage devices on the network. Whether it is a software-centric approach or a hardware-centric approach, we consider several fundamental elements – connectivity, security, development/deployment and data management and analytics – all of which must work in concert with each other for the Internet of Things to become a reality for the end user. We summarize them and later discuss about their advantages and challenges.

Whether it is a software-

Connectivity Connectivity between devices and the IOT platform can be established in multiple ways depending on the devices to be connected and solutions to be offered. Some devices connect to the IOT platform directly using protocols designed for IOT, some through existing wireless protocols, and others make use of protocol adapters provided by IOT platforms. Several protocols including MQTT, CoAP and AMMP are being developed, and most include the prerequisites for protocols in M2M and IOT such as low energy footprint, low coding footprint and the ability to function on heavily resource constrained devices.

development/deployment and

Deutsche Bank Securities Inc.

centric approach or a hardware-centric approach, we consider a several fundamental elements – connectivity, security, data management and analytics - all of which must work in concert with each other for the Internet of Things to become a reality for the end user.

Page 47

6 May 2014 TMT The Internet of Things

Another method of establishing connectivity between devices and IOT platform is through software modules installed on devices or on gateways connected to the devices. These modules are developed to be compatible with multiple operating systems. Overall, connectivity encompasses not just the protocol but how wireless networks leverage improved intelligence, dynamically leveraging resources through wireless networks and machine-to-machine communication, as well as how that technology could change in the future to adapt to more automated devices. Security Security will be paramount in almost every use case for IOT. If there are numerous compromised instances, then IOT may only realize a fraction of its potential. Securing both user data and the protection of connected devices against unauthorized access will be key. Ideally, security should be insured at all levels including the network, user, application, connection, system and database. This requires IOT deployments to incorporate security measures from the data center to the end user. Platforms, being central to deployment and hosting of applications and connections, play a vital role in the security of IOT.

IOT deployments must

IOT platforms will also need to minimize or eliminate the need for additional expenses to integrate developer’s security infrastructure with that of the IOT platform in order to support IOT applications and devices. In an ideal scenario, security features of IOT platform will work in tandem with existing security infrastructure at the client’s end to ensure integrity of the security system and data. This will increase the value proposition of IOT platforms than do-ityourself (DIY) approach taken by some developers.

In an ideal scenario, security

incorporate security measures from the data center to the end user

features of IOT platform will work in tandem with existing security infrastructure at the client’s end to ensure integrity of security system and data

Application development platforms typically consider security from the design phase and it is a never ending development process, even when end-users start accessing and interacting with the application. Besides the use of SSL certificates, various application platforms use Transport Layer Security (TLS) to encrypt message traffic and enhance security at the communications level. Security is enhanced by implementing various other procedures such as requiring multiple layers of authentication, restricted user rights and meticulous policy management, allowing only authorized users to be given rights to access the system. Sometimes authorized users are restricted access to data and actions they can perform, it just depends on how the security layers are developed. If this all sounds complex, with a web of overlapping requirements, that is because it is. Security is a critical element which must be designed-in from the start and tweaked on an ongoing basis in order to deal with the complex threats of a different landscape and changes that could be ahead. Development and deployment: Market research firms have different estimates of how many connected devices there will be in the next 10 years. Estimates range from 23 billion to 50 billion. Regardless of the estimate, the end number is a magnitude greater than where we stand today. The growth in connected devices is likely to be fueled by an explosion in applications, purpose built for areas ranging from health care to industrial systems to provisioning of utilities, running on devices that range from small to large, remote to local.

Page 48

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Similar to how frameworks like Java, ASP.NET, Ajax and others helped spur growth in web applications, dedicated application development platforms for IOT will play a critical role in creation, hosting and management of applications for Internet of Things. In general, application development platforms reduce the time and complexity involved in creating new apps. By leveraging these platforms, developers can not only minimize development efforts but also avoid incurring high costs. Figure 27: ThingWorx Application development platform

Source: ThingWorx

Some IOT platforms have become sophisticated enough to allow developers to create apps without writing a single line of code. Platforms like ThingWorx have even created an app store to aid development of applications on its platform. ThingWorx app store categories range from analytics to cloud services to enterprise systems. In the ThingWorx marketplace, vendors such as Google, Oracle, SAP, SalesForce and Ericsson offer various tools in the form of maps, database management, analytics and protocol adapters that a developer can simply embed in the application and leverage its functionality. To further accelerate application development, ThingWorx platform offers app templates for pre-defined use cases. In short, helping developers create applications in a relatively short period and manage them with minimum effort.

Some IOT platforms have become sophisticated enough to allow developers to create apps without writing a single line of code

Although most application development platforms primarily rely on the cloud for deployment, some platforms allow developers greater choice, giving them a choice to deploy applications on premise, in self hosted environments, on gateways, or simply on the device itself. That said, application development platforms are simplifying the process and making it more flexible at the same time. By doing so, the development time is shortened and the number of use cases that can be addressed with IOT technology is expanded.

Deutsche Bank Securities Inc.

Page 49

6 May 2014 TMT The Internet of Things

Data Management and analytics IOT will invariably result in large amounts of data, which is one of the intended outcomes. The next step is to leverage this data in order to assist businesses with decision making, customer experience, operational efficiency and potentially much more. This represents an evolution from earlier versions of M2M communication wherein connected M2M devices were primarily used for monitoring the status of assets, rather than gaining business insights using data management and analytics. This intelligence will not only be used for recording and transmission of data but also to filter noise from data so that only relevant information is recorded and sent to the datacenter or cloud. Platforms will likely play an important role in determining how data generated by sensors and devices are collected, stored, processed and analyzed. Data received can be in an unstructured form, needing to be organized and loaded onto enterprise data warehouses. Depending on the agility of platform, it could help developers gain real-time visibility into data using various analytics driven programs. New generation M2M/IOT platforms could also make it possible for data to be shared with multiple applications. In short, the ability to analyze data in a real-time manner in order to leverage it will be just as critical as the elements listed above. As such, we spend a considerable number of pages in this report discussing the possibilities of what can be done with an effective data management platform.

Platforms will play an important role in determining how data generated by sensors and devices are collected, stored, processed and analyzed. Data received can be in unstructured form, needing to be organized and loaded onto enterprise data warehouses.

IOT platform for devices The Internet of Everything is not just about applications and analytics. IOT is as much about the hardware and devices connected to the internet as it is about the applications running on them. Some IOT platforms will embed the software supporting their platform on the connectivity chips, thereby creating a tighter integration between the platform and devices. Thingsquare, Electric ImpOne and Ayla are some such platforms. Qualcomm has taken a hybrid approach to IOT. It is working with Oracle to bring Java enabled development platform to connect, locate and control everyday objects. The platform comes with an onboard sensor for light, temperature and movement, as well as embedded modules for connectivity via Wi-Fi, Bluetooth or cellular radio. Alongside this, Qualcomm is working to develop software standards via the AllSeen alliance, which will allow everyday devices like refrigerators, washing machines and televisions to communicate with each other as well as with people. To strengthen its viability as a universal platform, Qualcomm has brought in AT&T to test the applications and products on the platform. Developers can now test their products on AT&T’s network during varying stages of development. We discuss this in more granular detail in following sections of the report. In this framework Qualcomm has adopted a hardware plus software approach to its Internet of Everything development platform. Qualcomm offers a comprehensive development platform that supports Oracle Java embedded SDK and offers support for both hardware and application development. Devices designed on Qualcomm’s platform have in-built cellular connectivity. In addition, the presence of onboard sensors, indicators and relays help minimize development time of the product, thereby reducing time to market for new devices.

Page 50

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Qualcomm’s lead in SoC technology and its ability to incorporate multiple connectivity modules on single chipset eliminates the need to have a discrete microcontroller/processor or memory while allowing for multiple options to incorporate connectivity in the device, thereby reducing the size and the total bill of materials.

Qualcomm’s lead in SoC

To facilitate portability, Qualcomm allows the code to be ported to any QSC6270T Module with Java for the commercial product.

need to have a discrete

technology and its ability to incorporate multiple connectivity modules on a single chipset eliminates the microcontroller/processor or memory while allowing for

Figure 28: Qualcomm’s IOT development platform

multiple options to incorporate connectivity in the device

Source: Qualcomm

Broadcom is taking a different tack with a platform called Wireless Internet Connectivity for Embedded Devices (WICED). It simplifies the process of incorporating wireless internet connectivity in devices such as refrigerators and washing machines and provides connectivity using Wi-Fi, Bluetooth and Bluetooth Low Energy (BLE) modules. Utilizing Broadcom’s WICED platform, developers can create secure wireless networking applications on existing microcontrollers in the products and with their new SDK, Broadcom makes it easier for device makers to design speakers and audio devices to connect with Apple’s AirPlay. The WICED platform enables development of network connected applications that run on highly resource constrained microcontrollers like the ones with 512kB Flash memory and 64-128kB RAM. So Qualcomm is not alone here – many connectivity vendors are introducing IOT modules (platforms) into the space, thereby giving OEMs a turn-key solution to integrate into their products.

Deutsche Bank Securities Inc.

Page 51

6 May 2014 TMT The Internet of Things

Figure 29: Broadcom WICED module

Source: Broadcom

Of all the building blocks of IOT, a platform approach will likely have the widest and most extensive use given its ease of device integration and its ability to offer a turn-key solution that will encompass more functions than any other element of IOT alone. Currently, no single platform has gained widespread adoption or reached a tipping point whereby other device makers need to follow along simply because the consensus is building quickly. We are still in a place where different applications have different development, management and support requirements. Put another way, there is no universal solution that will work just as effectively for a home electricity meter as it does for a home video monitoring system. One of the alternatives to the adoption of platforms is the do it yourself (DIY) inclination of certain organizations. Though we have yet to find conclusive evidence on whether outsourcing IOT platform is actually feasible, it could be a cost-effective approach for small and mid size business, which need a custom solution for cheap. Large organizations will likely decide based on the importance of IOT applications to their core operations.

We are still in a place where different applications have different development, management and support requirements. Put another way, there is no universal solution that will work just as effectively for a home electricity meter as it does for a home video monitoring

Arduino Arduino is a platform intended for products that interact with environment sensors and respond by controlling lights and other appliances. The company sells preassembled boards that can either be used as is or be adapted based on user’s need. Given the flexibility and low cost, the solution has become relatively popular. Arduino also provides software which is compatible with Windows, Linux and Mac OS.

Page 52

system

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 30: Arduino microcontroller board

Source: Arduino

Intel Galileo Intel has developed a platform called Galileo. Launched in October 2013; it is a microcontroller board based on Intel’s 32-bit processor Quark SoC. Galileo leverages the simplicity of Arduino’s software and hardware environment and is the first board designed on Intel’s X86 architecture to be compatible with Arduino. Galileo represents an alternative to more complex Atom and Intel Core processor-based designs. Through PC based ports and features Galileo extends beyond the Arduino ecosystem. Figure 31: Galileo board

Source: Intel

Deutsche Bank Securities Inc.

Page 53

6 May 2014 TMT The Internet of Things

Are Networks Ready? Beneath all the connected devices, their respective applications and layers of analytics, is the network itself. Wireless and wired networks are the backbone of IOT. These networks will connect billions of devices (all very different from each other) and establish integration of physical world (end-devices) and digital world (servers, applications, cloud analytics). Legacy telecom networks have evolved in the past to support newer technology, services and applications. The process of network transformation, however, was slow until smartphones arrived. Since that point, we have seen a relatively faster transformation of networks from 2G to 3G to 4G and beyond. The emergence of IOT and resulting surge in data has enormous implications for telecom networks, enterprise networks and cloud infrastructure. Billions and Billions of devices No matter the third-party research firm, every firm agrees that there will be multiples of tens of billions of connected devices by 2020. To obtain meaningful information from these devices they need to be identified and addressed by the network. These devices will either connect directly to the network or via gateways. In doing so, networked devices will establish connection with other devices, humans and control systems. As connected devices make consumers more comfortable, businesses more efficient and governments more responsive, these stake holders will become ever more reliant on mobile networks for delivery of services. In this process significant value will be generated.

No matter the third-party research firm, every firm agrees that there will be multiples of tens of billions of connected devices by 2020

Figure 32: IOT brings value to businesses, consumers and assets

Source: Cisco

Present-day networks, though capable of delivering different applications and connectivity solutions, were not designed for wide ranging IOT solutions. IOT enabled networks ought to have provisions to carry a very large number of devices yet have the flexibility to offer customized solutions with respect to connectivity, cost and quality of service.

Page 54

Present-day networks, though capable of delivering different applications and connectivity solutions, were not designed for wide ranging IOT solutions Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

This flexibility is important due to the fact that different IOT applications will have significantly different requirements from the network; some applications require low latency, always-on connectivity and frequent data transmission while other applications can make do with intermittent connectivity and infrequent data logging. Ericsson has proposed a SaaS solution to this problem by introducing Ericsson Device Connection Platform (EDCP). It helps network operators to bring IOT devices on the network and support new applications by providing connectivity between devices and enterprise applications, enabling policy enforcement, controlling the quality of service for different applications and making provisions for future deployments. In short, EDCP allows operator’s mobile network to act as a flexible link between enterprise applications and remotely located IOT devices. Figure 33: Ericsson Device Connection Platform – general deployment

Source: Ericsson

Another important aspect of network and device readiness for internet of things is the ‘gateway’. With our discussion on gateways, we aim to address the issue of connecting disparate devices and legacy systems to enterprise data centers and cloud where relevant information can be processed and analyzed. Sensors and actuators are nothing new. They have been around for many decades, are used for specific applications, are installed using proprietary technology, have customized interface and run on specific protocols. Sensors are enablers of IOT; however, their fragmented deployment inhibits true IOT. Efforts to standardize protocols, platforms and device hardware to drive growth of IOT are underway. Once these standards are ratified, subsequent IOT deployments will have seamless integration and interoperability. However, a large number of existing deployments lack standardized components. Deutsche Bank Securities Inc.

Page 55

6 May 2014 TMT The Internet of Things

To extract optimal value from IOT, organizations need to overcome the issue of fragmented deployment. One possible path is to opt for new and standardized implementation, but that is a costly affair. Imagine if energy distribution companies were required to replace hundreds of million dollar transformers to realize the benefits of IOT. The other option is to establish integration using gateways that facilitate integration of legacy systems with new business applications and analytics. We believe the latter is the path that most businesses will take, optimizing the value of their existing business assets.

Gateways have traditionally been used to bridge dissimilar

Gateways have traditionally been used to bridge dissimilar technologies and integrate heterogeneous interfaces. They also enable integration of devices running on different protocols. Gateways can be used in a similar fashion with the internet of things to institute intelligent connectivity to legacy device networks. Gateways not only act as enablers of integration, but also as a protective layer between application and sensor. A resource constrained sensor may run out of battery if it is in direct contact with the application that constantly queries the sensor. The presence of a gateway prevents direct interaction between sensor and gateway, thereby extending the battery life of the sensor. Gateways can also cache and aggregate sensor readings so that the data log can be sent in one packet to the application rather than establishing frequent contact with the network, thereby bringing down the load on the network.

technologies and integrate heterogeneous interfaces..

…Gateways can be used in a similar fashion with IOT to institute intelligent connectivity to legacy device networks.

Figure 34: Gateways in the context of IOT

Source: Ericsson

With the progression in IOT, networks will adapt to the changing connectivity needs of devices and applications. Currently, smartphones, tablets and many other connected devices are always on and always connected. They communicate continuously with the network, often pinging the network in the background without us noticing it.

Page 56

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

In contrast, most IOT deployments should not require continuous network connection for data logging. In such a situation, one of the possible implementations of the network connection is to operate on discontinuous reception (DRX) mode. It not only brings down network load but also helps save power in energy deprived, low cost sensors nodes. Sensors will benefit by not having to remain connected at all times, thereby saving on energy requirements.

In contrast, most IOT deployments should not require continuous network connection for data logging. In such a situation, one of the possible implementations of the network connection is to

Based on current LTE specifications, the lengthiest possible DRX cycle is restricted to 2.56 seconds, which is very low compared to what many IOT applications can easily accommodate. Think of a temperature sensor in an orchid sending temperature data once every hour. Having DRX cycle of 2.56 seconds limits the energy saving quotient. Lengthening DRX cycles can help bring down energy consumption. According to Tuomas Tirronen of Ericsson Research, “using very long DRX cycles and optimized LTE procedures show that it could be possible to reach lifetimes of even over 10 years using off-theshelf and relatively cheap batteries” (http://labs.ericsson.com/blog/4g-for-iot).

operate on discontinuous reception (DRX) mode.

Figure 35: DRX cycle effect on battery life

Source: Ericsson

Increasing DRX cycles is in works in 3GPP Release 12, which is expected to be frozen in 2H14. If Release 12 specifies lengthening of DRX cycles in response to the need for lower power consumption it will be a reflection of changes in network in response to IOT.

Deutsche Bank Securities Inc.

Page 57

6 May 2014 TMT The Internet of Things

The rising Intelligence in Sensor Nodes The idea of IOT draws heavily on extracting data from wireless sensors. Sensors should be able to filter this data in a contextual manner and forward relevant instances of information to the gateway. Moving forward, data could and likely will be analyzed locally. Keep in mind, that some sensor deployments will be in harsh conditions, needing the ability to adapt, while others will operate in noisy (think congested) environments, requiring them to discern signal from noise. Smart algorithms, intelligent software frameworks and smart networking solutions can all help but these are very much in the early stages of development, and it will be some time before sensors will be able to leverage these in a meaningful way. Despite the timeline for intelligence at the edge, sensors are still proliferating across a number of verticals. Use cases It is easier to broadly categorize sensors into three categories – environmental monitoring, object monitoring and interaction monitoring. Environmental monitoring is defined as sensors measuring specific activities of the environment. Object monitoring measures the activity of both animate (humans) and inanimate objects (car). Interaction monitoring is the play between object and environments. Currently, there are a number of use cases being developed and some are already in use. The list below is certainly not comprehensive, but is meant to give readers some sense of the possibilities.

Page 58

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 36: A partial list of sensor use cases Cities

Parking Buildings safety

Monitoring the condition of buildings by sensing vibrations, material change

Noise monitoring

Monitor of sound in noise-restricted areas like hospitals

Traffic solutions

Solving traffic congestion issues by monitoring vehicle and pedestrian traffic and suggesting optimum routes

Street Lights Waste management Environment

Convey parking space availability in an efficient manner

Forest Fire Air pollution

Light sensors can adapt to the weather conditions to give best results Monitor the garbage level in containers so that garbage collectors can optimize their route Inform firefighters about fires. Sense drying areas and inform forest departments to take preemptive measures Monitor emission of CO2 and other harmful gasses in factories and cities to take preventive measures on time when the time required

Earthquake Landslide Water management

Metering

Water supply

Develop early earthquake detection model Monitor soil vibrations, moisture and earth density, so that high probability of landslide can be detected and concerned people are informed in time Check quality of water which is being supplied in the city

Leakage

Alert leakage of chemical waste of factories into natural water bodies like river

Pollution

Monitor the pollution level in the sea and other water bodies

Electricity

Monitor consumption of energy

Storage

Monitor the level of gas, oil and water in containers

Pressure Security

Retail and supply

Measure the pressure of water, air, etc

Access control

Detect unauthorized access in controlled areas

Detection

Detect release of dangerous gas or chemical in the environment

Tracking

RFID track location of products during transportation

Payments Intelligent Shopping Applications Industries

NFC payments enabled by use of smart sensors Development of intelligent applications to get price, deals and bundling

Monitoring

Monitor the level of airborne chemicals in a plant

Location Agriculture

Green house Irrigation

Home

Energy and water Detection

Health

Monitoring

Asset location in a factory Monitor green house effect; realize optimum green effect of the crops Monitor conditions and alert to exactly where water is needed and how much resulting in water conservation and better yield Effective use of energy, lighting and water by continuous monitoring Detect opening and closing of windows and doors of a house, enhancing the security of homes Monitor vital stats such as blood pressure, heart beat rate, temperature

Measurement

Measure UV ray density

Source: Deutsche Bank;Lbeliium

Remaining challenges Security, Quality of Service and Configuration remain the key challenges. Even in remote environments, sensors should maintain confidentiality, availability, integrity and authenticity. In our section on security we address the challenges and some possible solutions. In terms of Quality of Service, intelligent sensors are equipped with intelligent platforms with the goal to quality of service in spite of noise in the environment. There are a couple initiatives centered on this issue. Middlware Link Applications and Networks (MiLAN), developed by University of Rochester, has a goal to enhance QOS in light of noise. MidFusion is another middleware initiative; the solution allows applications to specify a description of their sensory requirements, which can be then used to discover and select the best sensors. We expect that more will follow, as QOS will become more important as people start to increasingly rely on the sensor feedback.

Deutsche Bank Securities Inc.

Page 59

6 May 2014 TMT The Internet of Things

Another challenge is the configuration of wireless sensors, especially as devices could dynamically come in and out of the range of a network. In the current scenario, the configuration of sensor nodes is taken care of by middleware/ gateways or by users manually. Sensors are being developed that will be able to automatically configure and administer responsibility themselves. We expect that as the development of sensors progresses this feature will be baked into the development from the foundation level. Conclusion Sensors are very much the heart of IOT. Currently there are challenges, but we believe as the idea of what can be done with them expands and as people increasingly rely on this data, pushing Metcalf’s law across the critical threshold, developers will find a way to move that intelligence to the edge. We believe this point will likely come in the near term, but that the possibilities of what can be done with such sensors are still in their infancy.

Page 60

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Frequency challenges Almost every device in the IOT concept will use some form of wireless communication. The extensive use of wireless communication by a growing number of devices within the same area will pose challenges. Specifically, it will challenge the quality of the network, the scalability of the elements and the reliability of all included. There are a number of solutions that can address the issue, including developing new radio and service architectures. Moreover, frequency spectrum allocation can be adjusted according to new channel requirements.

The extensive use of wireless communication by a growing number of devices within the same area will pose challenges. Specifically, it will challenge the quality of the network, the scalability of the elements and the reliability of

Radio Technology Given the constraints a number of IOT devices will be placed under, either because of energy efficiency or cost, a large number of devices will rely on a single radio. The efficient use of multiple frequencies that many have become accustomed to using a smartphone simply are not practical in an IOT environment. Regardless, IOT will increase the demand of Radio-Frequency spectrum.

all included

Current spectrum allocation, usage modes and radio technologies are not currently capable of handling the communication and data generated in IOT, or at least what people think about as a seamless IOT world. In terms of open frequency, there are not many open frequencies left. One is the vacated digital television 800MHz frequencies and recently in Europe, these were changed to mobile broadband. Television spectrum bands below 1GHz (TV white spaces) have the potential to work efficiently with license-exempt technologies. In past research we walked through a number of potential usage scenarios (please refer to From Start to Mobile Clouds and Beyond). The actual technology being used will vary, depending on how far and how quickly the information will need to travel. For almost 90% of the IOT applications that are currently being considered, most will use short range wireless technology like Bluetooth, Zigbee and WLAN. To get a sense of the range and throughput of various radio technology, we found the following chart helpful.

Deutsche Bank Securities Inc.

Page 61

6 May 2014 TMT The Internet of Things

Figure 37: Radio technology range and bit-rate

Source: Innovate UK

To address spectrum scarcity in IOT, the physical layer should be equipped with spectrum sensing techniques and advance reconfiguration ability. Sensing techniques such as matched filtering, cyclostationary feature detection and energy detection have been brought up as possible ways to improve efficiency.

Sensing techniques such as

Matched Filtering is one of the most efficient spectrum sensing techniques. It leverages identification to demodulate only approved signals and as a result, gives out a strong signal to noise ratio. However, it is complex, as there is a high dependency on prior knowledge of target signals and the cost is also high.

up as possible ways to

matched filtering, cyclostationary feature detection and energy detection have been brought improve efficiency.

Cyclostationary Feature Detection uses spectral correlation analysis to enable signal identification. The interpretation of modulated communication signals can be done via multiplexed sinusoidal waves (type of waveform) with periodicity, known as cyclostationary. The advantage of this method is that it eliminates the need of multiple antennas as various waves can be distinguished by using spectral correlation function. However, this technique also has high dependency on prior knowledge of target signals. In energy detection, a channel that is idle or busy is determined by comparing the preset threshold with the power level in the frequency. The advantage of this method is that it does not require prior knowledge of target signals. Media Access Control Spectrum related functionality is regulated by the media access network (MAC) layer, whereas reconfiguration related to spectrum conditions are made at the physical layer. The regulation can be done locally by individual nodes or by cooperation between entities. The mode and duration of sensing are two aspects of spectrum sensing by the physical layer, which needs to be Page 62

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

controlled by the MAC layer. There are two modes of sensing, reactive mode and proactive mode. Reactive mode helps conserve power by sensing the spectrum only when the data is to be transmitted. Proactive sensing is exactly what you might think, the sensing of spectrum at regular intervals. There is also something called cooperative sensing, which is created through the MAC layer and leverages entities of the physical layer involved in spectrum sensing activities. As these entities work in relatively different circumstances, by collecting data from all of them, it improves the quality of decisions. Network Layer As it pertains to IOT, the network layer ensures data network connectivity and efficient data transmission paths. The routing of data is done in a very opportunistic way in wireless sensor nodes network as the reliability of links is very low. At each node the forwarding conditions are analyzed and data hops are done when they can be done, which is different than what we might envision happens with typical internet traffic. Techniques like spectrum awareness can increase the parameters on which forwarding conditions are analyzed at each node for optimum hops. Specifically, spectrum aware mesh routing (SAMER) adds error rate and bandwidth to parameters on which data links are analyzed. Network integration can help as well, but as we learned earlier with protocols and others, it poses a number of challenges.

M87 developed a technique

An Austin based start-up, M87, has developed a technique that enables a cluster of m2m devices to talk to the network through just one good radio amongst the cluster, essentially daisy chaining off the good radio. We believe this can both simplify and lower the cost of deploying wireless, IOT networks while improving reliability.

the good radio.

that enables a cluster of m2m devices to talk to the network through just one good radio amongst the cluster, essentially daisy chaining off

Conclusion The bottom line is that there are a number of constraints on almost every element of the network currently, from frequency allocation to the MAC layer management and the wireline network itself. There are techniques which can help alleviate some of this strain, but it will take some time before these are optimized.

Deutsche Bank Securities Inc.

Page 63

6 May 2014 TMT The Internet of Things

Big Data and Analytics We have all the data from remote and not so remote sensors, actuators, mobile phones, tablets and other connected devices. Now what? Not being able to turn the collected data in to actionable insights will create an issue, and likely not make the investment worthwhile. If done well, the opportunities to gain meaningful insights are only bounded by one’s creativity. Big data and analytics are at the heart of IOT. All the efforts around developing standards, adjusting networks to IOT and deploying applications are ultimately directed towards realizing economic and social benefits. These benefits come in various forms such as efficient use of natural resources, minimal breakdown of machinery, increasing productivity of employees, better customer satisfaction, disaster management and on and on.

Big data and analytics are at

To achieve these benefits, stakeholders rely on Big Data and analytics tools to harness data generated by billions of IOT devices. It is nearly impossible for IOT to realize its potential and fulfill its promise if the hardware, software and services elements of IOT analytics are not in place.

realizing economic and social

the heart of IOT. All the efforts around developing standards, adjusting networks to IOT, deploying applications are ultimately directed towards benefits.

IDC breaks the Big Data industry into three broad segments: infrastructure, software and services. Infrastructure is the largest segment with a 45% share of Big Data revenue in 2013, while services and software accounted for 30% and 25%, respectively. Growth in Big Data is driven by sustained advancement in cloud-based Big Data services and the introduction of features, such as YARN, which makes Hadoop a multi-application framework. The further enhancement of security and privacy measures plays a critical role in making Big Data solutions enterprise class and ready for mass adoption. Overall, the Big Data ecosystem is evolving with industry players forming technical and non-technical partnerships. Figure 38: 2013 Worldwide Big Data Revenue by Vendor ($ mlns) Vendor

Big Data Revenue

Total Revenue

Big Data Revenue as % of Total Revenue

% Big Data % Big Data % Big Data Hardware Revenue Software Revenue Services Revenue

IBM

$1,368

$99,751

1%

31%

27%

42%

HP

$869

$114,100

1%

42%

14%

44%

Dell

$652

$54,550

1%

85%

0%

15%

SAP

$545

$22,900

2%

0%

76%

24%

Teradata

$518

$2,665

19%

36%

30%

34%

Cisco Systems

$295

$50,200

1%

72%

12%

16%

Juniper Total

$28 $18,607

$4,669 n/a

1% n/a

82% 38%

0% 22%

18% 40%

Source: IDC

For all its benefits, carrying out analytics on a vast set of data produced by IOT devices is a challenge. The operational costs associated with storage of rapidly growing, unstructured data is a key hurdle to most organizations. Data has been growing at an unprecedented rate, so much so that 90% of the global data today was created in past two years alone (IBM). Page 64

90% of the global data today was created in past two years alone (IBM).

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

In addition to storage, organizations need to manage and secure data, along with creating scalable infrastructure to meet future needs. These are significant challenges. Our research suggests that some of these challenges are being addressed by corporations like IBM and Cisco. Figure 39: How IOT will effect the Big Data landscape

Source: IBM

Along with an increase in the amount of data, IOT will bring about several changes in the velocity and variety of data gathered. Outside of IOT, there are not a great number of use cases where we see enormous amounts of unstructured data being generated, transmitted and processed in real time to obtain actionable alerts. Most of present day data processing involves batch processing of structured data.

Most of present day data

However, that may soon change. IOT deployments will lead to large number of applications in which vast amounts of data would require real-time processing to derive timely business insights and obtain actionable information. Collection and periodic batch processing of data might lead to a delay in obtaining relevant insights, providing analysis that was good yesterday, but not necessarily today.

Collection and periodic batch

One solution to this issue is stream processing, which enables real-time collection, visualization and analysis of large scale data arriving in high velocity. Stream processing works alongside existing enterprise resources without disrupting the systems and storage infrastructure already deployed. It allows organizations to analyze structured and unstructured data arriving in large quantities.

today.

If implemented well, stream processing not only facilitates large volumes of data to be aggregated and analyzed but also helps in maintaining data quality while lending itself to data auditing.

If implemented well, stream

To give you a perspective on the rate at which certain web/mobile applications generate data, Facebook generates more than one Tb of data every hour while Twitter generates one Tb of data every two hours4. Real-time analysis of this data help businesses identify which topics are trending in a given geography at

4

processing involves batch processing of structured data.

processing of data might lead to a delay in obtaining relevant insights, providing analysis that was good yesterday, but not necessarily

processing not only facilitates large volumes of data to be aggregated and analyzed but also helps in maintaining data quality while lending itself to data auditing.

https://www.ibm.com/developerworks/library/bd-streamsintro/

Deutsche Bank Securities Inc.

Page 65

6 May 2014 TMT The Internet of Things

a given time among a given age group of people. It offers them monetization opportunities based on consumer interest and behavior. In the absence of stream processing, data would be collected, stored and later analyzed. By the time insights are obtained it might be too late to reach out to users as they may have moved on to something else. This is just the beginning; as IOT matures, application developers will come up with hundreds and thousands of applications that generate vast amounts of structured and unstructured data, requiring real-time analysis for efficient functioning. Think of a transit application that provides real-time information on busses and trains, real-time traffic updates or availability of parking in a given business district or shopping mall, while connecting it to current and forecasted weather data to help us decide whether to drive to our destination or use public transportation systems. Such an application solves a major, real world problem and requires real-time feeds that need real-time processing – sort of an extension of HopStop. Multiply this by millions of users and hundreds of cities for which such information is made available and you have a massive challenge at hand. Such an application will depend on stream processing and cannot be subject to batch processing. Figure 40: Stream processing enables real-time response

Source: EMC

Stream processing does have its challenges though. It requires extensive computing resources and needs to overcome hurdles related to security, bandwidth, capacity and data integrity. In addition, a vast amount of data arriving through IOT applications needs to be continuously filtered. Data will need to be filtered in the context of the application, as a one size fits all solution for filtering data does not work. All of this requires very powerful processors. Moreover, stream processing will significantly increase workloads in data centers; we estimate that a growing number of use cases of stream processing coupled with the ensuing complexity of the process will have a transformative impact on data centers. Existing data centers need to evolve, while new data centers should be designed to support the increased expectations for performance. Currently, building a new data center or upgrading an existing one in order to execute stream processing is a cost prohibitive and daunting task. Page 66

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

One of the solutions proposed to make stream processing commercially viable is to implement distributed data center management. To prevent individual data centers from being bogged down by a flood of IOT data, data center managers can decide to distribute data aggregation, tasking it to multiple data centers. These data centers can then perform the initial level of processing and send relevant data to the central data center for further processing and analytics. In addition to distributed data center management, data center managers need to leverage DCIM (Data Center Infrastructure Management). DCIM is a sophisticated data center monitoring solution which helps manage and optimize the performance of data centers. Today’s data centers are designed to accommodate bandwidth requirements originating from humanapplication interaction. A typical data center is not designed to handle massive sensor data, creating a wide gap in available and required bandwidth. To address this issue, data center managers will need capacity management platforms like DCIM.

Data center managers need to leverage DCIM (Data Center Infrastructure Management). DCIM is a sophisticated data center monitoring solution which helps manage and optimize the performance of data centers

DCIM gives a holistic view of the functioning of data centers by providing information about real-time utilization levels, energy efficiency, capacity planning and asset management. This information results in better control of data center assets and greater operational efficiency, allowing data centers to be able to take on larger workloads associated with stream processing and Big Data. DCIM brings down operational expenses by reducing data centers downtime. And based on recent commentary from industry analysts, modern DCIM solutions are sophisticated enough to drive utilization levels high enough to be able to defer data center capital expenses. Figure 41: iTRACS DCIM to optimize data center for stream processing

Source: CommScope

In short, IOT aims to make institutions and society more efficient; since IOT deployments generate vast amounts of structured and unstructured data that require efficient collection, aggregation and detailed analysis, Big Data Deutsche Bank Securities Inc.

Page 67

6 May 2014 TMT The Internet of Things

solutions will gain traction. Challenges still remain, in the name of velocity, volume and veracity of IOT data. It is likely that many IOT related Big Data problems are best addressed using stream processing, which is a daunting task and places strain on data center resources, requiring data centers to be managed at high levels of efficiency. DCIM is an effective way to optimize performance of data centers and to maximize their utilization levels. We would note that CommScope is one of the leading companies in DCIM space. Figure 42: The path to increased investment in DCIM

Source: Deutsche Bank

Page 68

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

IOT Use Cases In this section, we highlight 3 IOT use cases – representative of ongoing IOT buildouts in 3 major industry sectors: 1) Industrial Cyber Security, 2) Smart Energy (Oil and Gas sector) and 3) Smart Retail.

We highlight 3 IOT use cases – representative of ongoing IOT buildouts in 3 major industry sectors: 1) Industrial

Our use case discussions, albeit at a high level, are our basis for highlighting the investable plays in the “infrastructure” aspects of IOT buildouts – i.e. in the Edge Cloud, Datacenter Cloud and Network Intelligence and Security aspects of an IOT infrastructure.

Cyber Security, 2) Smart Energy (Oil and Gas sector; and 3) Smart Retail

We believe key catalysts for capex investments in IOT infrastructures are the potential for structural multi-year improvements in productivity and opex metrics. A recent GE IOT study notes potential for appx +$18B in annual opex savings benefits from a modest 1% improvement in operational efficiency metrics. The GE IOT study is basis for our view that capex investments in IOT infrastructures is likely to be pegged at around 10% of the $18B in yearly opex savings – i.e. we anticipate IOT related capex investments in the $2B range a year over the next few years on Layer 0/7 networking solutions. We note that CxOs typically invest ahead of anticipated recurring opex savings.

Figure 43: Industry data summary on CY15 network equipment capex spending [Infonetics] and DB view of incremental network infrastructure spending on IOT use cases

The $2B a year in IOT related networking capex investments would be incremental to the current +$47B a year of capex spent on networking equipment – summarized in Figure 43 – and a fundamental basis for potential upside to CY15/16+ consensus expectations.

Appx CY15 IT Capex Spend ($B; based on Infonetics data) $9.0

In the remainder of this chapter, we discuss 3 major IOT use cases and highlight how each of the Layer 0/7 IT stacks are involved in implementing the end to end IOT infrastructure. The networking vendor plays are in one or more of the IT stacks’ – as we summarize in the IOT use cases discussion. The basis for our IOT use cases discussion is as follows. 1.

2.

The IOT CyberSecurity use case highlights how IOT enabled infrastructure can offer a well-rounded defense against a broad range of security attacks – from the Operational Technology elements – i.e. physical security access, etc – to the Information Technology elements – i.e. Layer 0/7 network equipment. The Smart Energy scenario is representative of an IOT use case for Remote Monitoring and for Big Data Analytics of Gigabytes of data output from Internet connected sensors in offshore Oil and Gas rigs. From a network impact POV, the Smart Energy use case is “bandwidth intensive” and “delay insensitive”.

Product Category Datacenter Switching Enterprise Routing Carrier Routing & Carrier Ethernet

$4.0

$15.0

Layer 4/7 (Application Delivery; WAN Opt)

$3.0

Security Equipment

$7.0

DWDM Optical Networking

$9.0

Total Network Eqp Capex View for CY15 DB view of incremental IOT related networking capex spend

$47.0

$2.0

Source: Deutsche Bank; Infonetics

3.

The Smart Retail example is representative of a use case for IOT enabled supply chain management and Big Data Analytics. From a network impact POV, the Smart Retail use case is “bandwidth tolerant” and “delay sensitive”.

Deutsche Bank Securities Inc.

Page 69

6 May 2014 TMT The Internet of Things

In the discussion below, we discuss how the networking vendors can benefit from selling their policy-based and dynamically managed Layer 0/7 equipment in these 3 representative IOT use cases. IOT Use Case #1: Industrial Cyber Security Figure 44 illustrates a representative IOT architecture for implementing an industrial cyber security use case. The IOT architecture has both IT and OT network elements – required to implement an end to end cyber security solution. The OT aspects include industrial security systems and devices at the physical security layers – i.e. Internet connected objects such as programmable video cameras, audio recording devices, temperature sensors, fire alarm equipment, access badges, etc. The IT aspects are of interest in the context of our FITT – since they correspond to investable opportunities for the Layer 0/7 networking vendors in our coverage universe. Drilling down into the IT layers, we highlight security capabilities required at every major tier of the IT stack – i.e. at the optical transport and wireless access layers, at the Layer 2/3 switching and routing layers, and at Layer 4/7. Referring to the cyber security architecture in Figure 44, we note that vendors such as Cisco, Aruba, etc currently offer wireless air interface security and wireless data layer encryption mechanisms in their industrial hardened 802.11ac access points.

Page 70

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 44: Illustration of IOT Cyber Security network architecture

Applications

Enterprise c Apps

Big Data c Analytics

Datacenter Cloud

Datacenter Cloud

IOT c Control

App Specific QoS; Big Data Analytics; Traffic & Policy Management (F5, Cisco ACI, VMware, IBM…)

Layer 3 Leaf / Spine Switching Network & Virtual Overlays (Cisco Nexus 9k, bare-metal switches, etc) Application + Network Security; DNS (F5, Infoblox, Cisco, etc)

Assumes Variable-Latency; Variable BW Network Link

Edge Cloud Node

Edge Cloud

Enterprise Edge Router w/ Local Compute, Networking, Storage, Security

c

(e.g. Cisco ISR 819; HP MSR; OEM Edge Routers)

IOT End-Points

IOT End-Points

Wireline & Mobile IOT End-Points; IOT Sensors, Actuators, Mobile Devices, etc

Point of Sale Sensors; Digital Signage, Video Cameras, Sales Metrics Analyzers

Source: Deutsche Bank

Cisco, with its +50% market share in enterprise WiFi equipment and broadest set of networking solutions in enterprise IOT use cases, is likely to be a mid- to long-term beneficiary of enterprise WiFi networking rollouts in IOT architectures. Cisco’s point product competitors, we note, lack the IOT solutions integration expertise that Cisco and its industrial IT ecosystem partners (Rockwell, GE, Siemens, etc) would bring to industrial IOT use cases. Moving up the IT stack, to Layer 2/3 switching and routing, we see Cisco likely to benefit in the CY15/16+ timeframe from sales of virtualized network firewall modules at the Edge Cloud Nodes (e.g. virtualized intrusion detection and

Deutsche Bank Securities Inc.

Page 71

6 May 2014 TMT The Internet of Things

prevention SW modules running in the Cisco’s ISR 819 series enterprise router, etc) and from sales of campus switches (e.g. Cisco’s 3850 series) running 802.11ac access points integrated into the switch. Moving further up the IT stack, into the Datacenter Cloud Layer, we note meaningful opportunities for Layer 4/7 specialists such as F5 and Infoblox to benefit from the IOT Cyber Security use cases. F5 – a market leader in Layer 4/7 has multiple security touch points in the IOT Cyber Security architecture that we illustrate in Figure 44. At the network access control layer, F5’s Access Policy SW Module – running as either a virtual instance or in the company’s HW optimized BIG IP or Viprion platforms – has an important role to play in controlling end-user access to a wide range of datacenter application servers – at a fine level of granularity – i.e. access control by user name, workgroup affiliation, geographic coordinates, user context, etc. IOT sensors imbedded in various industrial objects can relay security incidence monitoring events to F5’s Access Policy Module – and F5’s TMOS SW is designed to cross-correlate across 1000s of individual security incidence response and access control events – to generate security trend analysis patterns, feedback to IT security policy design, etc. We note that F5’s Access Policy Module is currently the best selling SW module in the company’s SW portfolio – a point of proof of the F5’s best–inclass competency in network and application access policy management. At the application and network firewalling layers, F5 has a core set of value propositions to offer – as part of an overall industrial cyber security framework. F5’s Application Security and Network Firewalling SW Modules – implemented as either virtualized SW instances or running in the BIG IP or Viprion HW platforms – can scale to 100s of Gigabits of packet processing throughput (in comparison, we note that many of F5’s security peers scale to 10s of Gb/s) for handling volumetric attacks such as Distributed Denial of Service (DDOS) attacks – originating from the public Internet – and targeted towards the perimeter of the extended enterprise (i.e. the enterprise’s HQ network, branch offices, key supply chain partner networks, etc). More recently, F5 has been offering a Cloud based subscriptions service (leveraging the Versafe acquisition) for detecting and IT reporting of advanced malware threats across IP connected devices – from mobile devices running industry standard OS to industry vertical devices running Linux or Windows OS. F5’s Cloud Security IT services – competitive with Cloud services offerings from Cisco/Sourcefire, FireEye, Akamai, etc – are likely to see directionally improving enterprise IT customer adoption trends in IOT and other industry vertical use cases – given our view that IOT, in particular, involves the provisioning of security capabilities for1000s of Internet connected end-points (in a wide range of form factors) – especially to defend the enterprise IT infrastructure against a growing number of advanced malware threats and sophisticated security attacks targeted at layers 4/7.

Page 72

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

F5’s iRules scripting framework has a core value proposition in IOT use cases – given the relative ease at which simple English language IT and business security policies can be defined and implemented in a matter of minutes – to defend the enterprise architecture against a dynamic set of security threat vectors, which by definition need a dynamic policy creation and implementation framework (offered by iRules). We estimate F5’s application and network security opportunity in the +$2B a year range (out of a total +$8B security equipment market [Infonetics]) – with growth rate trends in the low double-digit Y/Y range. From a security attach rate POV, it is our view that F5’s Layer 4/7 systems (BIG IP and Viprion) could see the security SW modules attach rate scale from the current +30% of HW platforms sold to around +50%, over the next few years. We believe IOT use cases are a key catalyst for scaling the company’s security solutions sales opportunities given the significantly higher number of security IT end-points in IOT use cases, which could be the underlying basis for scaling the quantity and types of security modules required to defend against a broad set of attack vectors in IOT Cyber Security use cases. Infoblox – a market leader in DNS and IP Address Management [IPAM] solutions – could secularly benefit from rollouts of DNS Security mechanisms to safeguard DNS servers situated at the enterprise’s infrastructure borders, i.e. at the network perimeter that connects the enterprise network to the public Internet and to the company’s’ supply chain partners. We note that a vast majority of enterprise DNS servers (which resolve incoming Web requests such as www.example.com to an set of IP addresses; corresponding to internal servers) currently lack adequate security mechanisms to 1) safeguard the DNS servers from security attacks that originate from the public Internet (e.g. DDOS attacks targeted at corporate DNS servers) and 2) defend against security attacks that originate from malware-infected employee PCs, smartphones, tablets, etc, when the employee computing devices are connected to trusted corporate networks. Infoblox recently introduced DNS Security appliances that defend against DNS security attacks that originate from the “outside-in” and the “inside-out” – and these DNS security appliances leverage live DNS reputation feeds to crosscheck all incoming and outgoing DNS queries against a constantly updated live IP address and Web URL feed so as to identify malicious URLs and hacking attempts on a company’s DNS servers. We estimate Infoblox’s DNS security opportunity for enterprise IT use cases (including IOT use cases) in the mid $100s of million a year range given our view that DNS security is a next-gen security use case and the market opportunity for Next-Gen Security Figure 45 is an IT stack level diagram summarizing how an IOT enabled Cyber Security architecture can be implemented by utilizing best-in-class security mechanisms at every major layer of the IT stack.

Deutsche Bank Securities Inc.

Page 73

6 May 2014 TMT The Internet of Things

Figure 45: Illustration of key IT stacks involved in implementing an IOT Cyber Security architecture

Application Security

Transport Layer Security

Network Layer Security

Application Access Control, Web Application Security, DNS (e.g. F5, Infoblox, etc)

Transport Layer Security (SSL) (e.g. F5, Citrix, A10, etc)

Virtual Network Layer Security (e.g. VMware, Cisco, Open Source SW, etc) Network Firewalls (e.g. Cisco, F5, Palo Alto, Juniper, etc)

IOT Endpoint Security

IOT Endpoint Security (Access Badges, Video Cameras, IOT Sensor Metrics, etc)

Source: Deutsche Bank

IOT Use Case #2: Smart Energy (Oil and Gas) Figure 46 illustrates a representative Edge Cloud and Datacenter Cloud architecture for implementing an IOT Smart Energy use case – in this case, a set of Internet connected sensors and actuators in offshore Oil rigs – for implementing remote IOT enabled systems and devices monitoring. Remote monitoring is a core value proposition that leverages centralized onshore facilities for dynamically managing a set of Internet connected endpoints in offshore Oil and Gas installations. A set of IOT objects (industrial sensors monitoring the entire lifecycle of offshore oil rig equipment; actuators implementing various control decisions, etc) continuously transmit sensor data (coded using industrial SCADA protocols and transported over TCP/IP) to Edge Cloud Nodes (e.g. Cisco’s ISR 819 router, etc) via wired Ethernet or WiFi LAN access. The Edge Cloud Nodes offer a core set of localized compute, networking, storage and security functions for the offshore oil rig sensors and actuators. The Edge Cloud functions include: 1) aggregation of IOT events from 100s or (1000s) of Internet connected sensors, 2) real-time local SW processing of IOT event data, 3) implementation of edge security access and intrusion detection capabilities and 4) batch upload of Gigabytes of IOT offshore rig data, etc.

Page 74

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 46: Conceptual view of an IOT architecture to support a Smart Energy use case

Applications

Big Data c Analytics

ERP c Apps

Datacenter Cloud

Datacenter Cloud

IOT c Control

App Specific QoS; Big Data Analytics; Traffic & Policy Management (F5, Cisco ACI, VMware, IBM…)

Layer 3 Leaf / Spine Switching Network & Virtual Overlays (Cisco Nexus 9k, bare-metal switches, etc) Application + Network Security; DNS (F5, Infoblox, Cisco, etc)

Assumes Variable-Latency; Variable BW Network Link

Edge Cloud Node

Edge Cloud

Enterprise Edge Router w/ Local Compute, Networking, Storage, Security

c

(e.g. Cisco ISR 819; HP MSR; OEM Edge Routers)

IOT End-Points

IOT End-Points

Wireline & Mobile IOT End-Points; IOT Sensors, Actuators, Mobile Devices, etc

Drilling Site IOT Sensors, Actuators, etc

Source: Deutsche Bank

The Edge Cloud nodes perform an important networking function; i.e. they transmit Gigabytes of IOT event data, aggregated from multiple IOT sensors in “batch” mode, to centralized databases and application servers. By batch mode, we refer to data sets being set on a “best-effort” bandwidthavailable basis to centralized datacenter servers versus in a continuous realtime mode. Deutsche Bank Securities Inc.

Page 75

6 May 2014 TMT The Internet of Things

We note that transmission of bulk data along with delay-sensitive IP traffic in a narrow bandwidth onshore to offshore WAN link often negatively impacts time-sensitive IP voice and data communications. This is the basis for our view that bulk data transport in WAN links needs to be dynamically managed to ensure overall integrity and performance of IP voice and data communications between onshore and offshore Oil and Gas facilities. Cisco’s ACI Fabric (anticipated to be generally available mid CY14 timeframe) could be relevant in Smart Energy networks given our view that network equipment in the WAN links (e.g. Edge Cloud Nodes running Cisco’s APIC Enterprise SW module) can dynamically communicate the bandwidth and latency requirements of application traffic traversing the links between the edge of the network and the datacenter to a centralized SDN controller (e.g. Cisco’s APIC SW engine). The SDN controller, in turn, can provision relevant Quality of Service policies on a box by box basis – i.e. in every switch or routers across the end to end network connection – so as to ensure optimum delivery of delay sensitive and bulk data traffic through the WAN links that interconnect the onshore and offshore Oil rig facilities. Figure 47 illustrates the relevant IT stacks involved in the dynamic management of the onshore to offshore network connections. Figure 47: Conceptual view of relevant IT stacks involved in the Smart Energy use case

Application Layer

Datacenter Cloud Layer

Edge Cloud Layer

Web 2.0 and SaaS Apps

Application Aware Networking, Traffic Management, Security (e.g. Cisco ACI Fabric, F5, Infoblox, OEM/Open Source systems)

Virtual Network Overlays (e.g. VMware, Cisco, Open Source SW, etc) Edge Routers w/ Local Processing & Security (e.g. Cisco ISR, HP, OEM edge routers, etc)

IOT Endpoints

IOT Endpoints (Drilling Site Sensors, Actuators, Analyzers, etc)

Source: Deutsche Bank

Page 76

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Referring to the IT stack diagram in Figure 47, we note that Cisco’s ACI solution would play a meaningful role in Quality of Service management at Layers 2/3, while F5’s policy management solutions could play a role for managing Web traffic at Layers 4/7. Cisco’s ACI Fabric (competitive with VMware’s NSX solution, etc) can also be used to set up virtual network tunnels between WAN router end points and datacenter servers – or between various server racks within the Datacenter Cloud – for managing the data traffic in the Smart Energy infrastructure based on policy descriptors such as virtual tunnels dedicated for routine backups, disaster recovery, delay sensitive traffic, best effort bulk data transport, etc. F5’s Web traffic steering feature – a core traffic management feature in the BIG IP and Viprion ADCs – has a core value proposition to offer for managing Web traffic requests originating from offshore oil rig installations. Different sets of Web URL requests – based on the business sensitivity and time-sensitivity of the application traffic – can be directed to different sets of Web application servers by F5’s ADCs running the core traffic management software (TMOS SW). F5 also offers iRules scripts to selectively encrypt different sets of Web application traffic, and also safeguard application traffic from eavesdropping and hacking attempts at the Web layers. F5’s Network Access Control solution (based on the Access Policy SW module) can be used to enforce selective tiers of access to energy management and data analytics applications in the Datacenter Cloud. As we noted in the Cyber Security use case discussion, Infoblox’s IP address management solution and DNS’ security solution have a key role to play in the Smart Energy use case: for automating IP address assignment across device types (wired and mobile end-points, servers, etc) and for safeguarding the corporate DNS infrastructure. We also note opportunities for Ciena’s Carrier Ethernet solution and optical network management software tools in the Smart Energy use case. Undersea cable – carrying optical Ethernet data traffic – is a relevant use case for sales of Ciena’s carrier Ethernet switches at either ends of the offshore and onshore optical WAN links. Ciena’s optical network management SW could be used to dynamically provision optical network transport equipment at either ends of an optical fiber link. In summary, we note near-term opportunities for the Layer 4/7 specialists – F5, Infoblox, etc – to sell into network intelligence and security use cases in the Smart Energy use case, opportunities for Ciena in optical Ethernet and optical network management tools and out-quarter opportunities for Cisco’s ACI Fabric solution. IOT Use Case #3: Smart Retail Figure 48 illustrates representative Edge Cloud and Datacenter Cloud architecture for implementing an IOT Smart Retail use case. A Smart Retail IOT use case is the converse of the Smart Energy use case from a networking POV. While the Smart Energy use case involves transport of Deutsche Bank Securities Inc.

Page 77

6 May 2014 TMT The Internet of Things

bulk data from the offshore rigs, the Smart Retail use case involves the dynamic management of mostly time-sensitive transactional data (e.g. inventory levels, hourly or daily product sales metrics, etc) from retail point of sale IOT sensors to regional offices and HQ locations of the corporate enterprise and its global supply chain partners. Referring to the IT stack level diagram in Figure 49, we note a variety of data traffic formats that need to be dynamically managed to and from the regional and HQ offices and the retail point of sale locations. At Layers 0/1, Operational Technology [OT] data such as parking space sensor data, motion sensor data, environmental metrics, etc is carried over Internet protocols (TCP/IP) to the Edge Cloud Node – where location-specific decisions can be made (often in near real-time) by SW applications that process the OT data in Enterprise Routers (e.g. Cisco’s ISR 819 series). At Layers 2/7, Information Technology [IT] data such as product inventory levels, customer demand trend metrics, product sales metrics by time of day or day of week, etc need to be delivered, often in near real time, to centralized Datacenter Cloud servers for business planning and decision making across the extended retail supply chain.

Page 78

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Figure 48: Conceptual view of an IOT architecture to support a Smart Retail use case

Applications

Supply Chain c Management

Big Data c Analytics

Datacenter Cloud

Datacenter Cloud

IOT c Control

App Specific QoS; Big Data Analytics; Traffic & Policy Management (F5, Cisco ACI, VMware, IBM…)

Layer 3 Leaf / Spine Switching Network & Virtual Overlays (Cisco Nexus 9k, bare-metal switches, etc) Application + Network Security; DNS (F5, Infoblox, Cisco, etc)

Assumes Variable-Latency; Variable BW Network Link

Edge Cloud Node

Edge Cloud

Enterprise Edge Router w/ Local Compute, Networking, Storage, Security

c

(e.g. Cisco ISR 819; HP MSR; OEM Edge Routers)

IOT End-Points

IOT End-Points

Wireline & Mobile IOT End-Points; IOT Sensors, Actuators, Mobile Devices, etc

Point of Sale Sensors; Digital Signage, Video Cameras, Sales Metrics Analyzers

Source: Deutsche Bank

We see a role for Cisco’s ACI Fabric solution (running on the enterprise routers and also in datacenter switches, etc) to dynamically manage the end to end network connections for reliable and express delivery of key performance metrics such as product inventory levels, product sales metrics, marketing promotions, etc between the retail locations and the extended supply chain.

Deutsche Bank Securities Inc.

Page 79

6 May 2014 TMT The Internet of Things

F5’s Layer 4/7 ADC solutions have a key role to play in the Smart Retail use case. F5’s iRules scripts can be set up to provide differentiated Quality of Service for in-store customers or in-store staff making secure (e.g. https://www.retail-store.com) purchases via the company’s e-commerce Web servers. F5’s core ADC software and specialized security SW modules can be utilized to defend the extended retail supply chain against sophisticated Web application layer attacks (malware in Web links, phishing attacks) and to handle unexpected surges in retail store Web traffic volumes to Datacenter servers, etc. Similar to our discussion in the Smart Energy use case, Infoblox’s IP address management solution and DNS security solution have a key role to play in automating IP address assignment across wired and mobile devices and in safeguarding the corporate DNS infrastructure against DNS layer threats. Figure 49: Conceptual view of relevant IT stacks involved in Smart Retail use case

Application Layer

Datacenter Cloud Layer

Edge Cloud Layer

Web 2.0 and SaaS Apps

Application Aware Networking, Traffic Management, Security (e.g. Cisco ACI Fabric, F5, Infoblox, OEM/Open Source systems)

Virtual Network Overlays (e.g. VMware, Cisco, Open Source SW, etc) Edge Routers w/ Local Processing & Security (e.g. Cisco ISR, HP, OEM edge routers, etc)

IOT Endpoints

IOT Retail Point of Sale & Supply Chain Endpoints (Point of Sale Sensors, Metrics Analyzers, Digital Signage, etc)

Source: Deutsche Bank

Page 80

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Companies Qualcomm Qualcomm has been positioning itself as a growth driver of IOT since cellular wireless technology first gained traction. The company started with Omnitracs, a platform for truck fleet management and from there the company branched out into a number of IOT initiatives. It developed the AllJoyn protocol, which is now overseen by the Linux foundation and has more partners and contributors than any other protocol platform being considered. We discussed AllJoyn in our note, as well as its IOT platform. These have both been used to develop revenue generating products in a number of industry verticals. Automotive: In terms of an overall platform, the company has only recently demoed a platform at Consumer Electronics Show and later at Mobile World Congress. The company retrofitted both a Jeep and a Mercedes, which both had multimedia streaming, connected device sharing, rear seat entertainment systems, enhanced navigation, vehicle diagnostics and location services. The platforms leveraged Snapdragon 602A processors, their connectivity solutions (Vive) and 9x15 modems. Qualcomm took the modems a step further and recently introduced the Gobi 9X30, with Cat 6 capabilities, Wifi hotspot and telematics services – a 20 nanometer modem specifically designed for the automotive category. It also includes vehicle to vehicle communication, a safety measure recommended by the National Highway Traffic Safety Commission. There are already plans for automakers like Audi and GM to use these solutions across a broad portion of their fleet and AT&T has partnered with Qualcomm to offer LTE connectivity services for cars. The GSMA estimates that there will be over 35 million embedded connectivity solutions by 2018. And while that may not be hugely meaningful to a company that will very shortly be manufacturing over 1 billion ASICs units per year, it will help accelerate consumers’ ideas of what having connectivity in other places outside of the smartphone can mean.

Figure 50: Qualcomm’s IOT initiatives AllJoyn

Protocol allowing elements to connect, agnostic of brand or operating system

IOT development platform Automation and Enterprise solutions

Platform for cellular connectivity Solutions that scale for enterprises

AllPlay media platform Plug and play wireless audio build on AllJoyn Smart Car

Telematics and computing designed for the car

Smart home and security management Wearables

Displays, efficient computing

Source: Deutsche Bank; Qualcomm

Figure 51: GSMA unit estimates for the connected car 40

Annual unit sales (million)

35 30 25 20 15 10 5 0 2009

2010

2011

2012 Embedded

2013 Tethered

2014

2015

2016

2017

2018

Smartphone

Source: Deutsche Bank; GSMA

Deutsche Bank Securities Inc.

Page 81

6 May 2014 TMT The Internet of Things

Smart home: At Mobile World Congress this year, Qualcomm demoed connected home solutions: everything ranging from a wine refrigerator, the door of which would trigger an alarm if left open, to remote locks, appliance alerts that pop up on the TV and even content monitoring for kids. Some believe that this market (connected home cellular devices) will quickly grow to just shy of 140 million connected units, from 25 million where it stands today. We are a bit more conservative than this; while many will quickly understand the benefits of adding connected solutions in their home, unless it is a seamless installation procedure or it is new home construction, it could take a bit longer. Many devices in the home will be connected via WiFi or Bluetooth. This will offer a cheaper way to leverage the internet, and will also offer quicker adoption given it can easily sync with the home gateway. To be clear, Qualcomm is a large player here as well, given its Atheros unit, but the ASP on the chipsets are much lower. Overall, we believe the home will be one of the earlier silos to become increasingly automated and adopted. It will likely be done early with unlicensed connected solutions and we have already seen this, but there are areas that will need to remain connected via cellular (e.g. smart meters and security systems). All of this will take time, but we believe Qualcomm is one of the best positioned companies to leverage the growing smart home trends.

Overall, we believe the home will be one of the earlier silos to become increasingly automated and adopted

Figure 52: Unit cellular estimates for the connected home

Source: Berg Insight; Beecham Research

Industrial and Enterprise solutions Qualcomm has developed cellular solutions for ATMs, POS, retails kiosks, digital signage, asset tracking and handheld terminals. The ease of set up, coupled with the ability to be always connected and mobile makes these devices appealing. While we do not think these unit volumes are meaningful to Qualcomm, we do think tablet attach rates are improving because of their increasing use as a point of sales mechanism. And while many will be connected via WiFi to the business’s network, a majority will have cellular backup given the importance of up-time on this device. The company also has a number of gateways and router solutions, which can be embedded into any device, from utility meters to card access solutions. The cellular element of these solutions negates the need for wire line, the installation of which can be a challenge. Given the security elements necessary for the enterprise vertical, we believe propriety platforms will remain for some time and thus limit overall unit growth here.

Page 82

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Qualcomm Development Platform Please see our section on IOT as a platform. Wearables Last year Qualcomm introduced the Toq, a smartwatch with its Mirasol display technology. The move was an attempt to prime the pump for wearable technology. While OEMs like Samsung and Sony have released their wares, the unit volume has not been stellar. These are early days for wearable technology. And the connectivity technology included in these devices likely will be of the unlicensed kind (WiFi or Bluetooth). There are a number of other ways Qualcomm can participate here, with the sale of their wireless charging technology, and to a lesser degree their display technology (though Mirasol has always struggled) and connectivity solutions. While we think wearables will eventually be a meaningful market, we think the technology has some way to go before it becomes a considerable number for component suppliers. Parting thought Qualcomm will be a key beneficiary of IOT; that is not an extraordinary statement. But given our thesis that it will take time for the market to develop, and with a company that is on the cusp of shipping 1 billion ASICs per year, it will be some time before it is significant for it . Having said that, when IOT unit volumes begin to ramp in earnest, Qualcomm will likely be one of the best positioned companies in the space, as it will have generated significant “lockin” given it has developed much of the foundational elements that drive even the most peripheral of areas of IOT.

Cisco The basis for our constructive view on Cisco in the IOT infrastructure opportunity in the out quarters (2H15+) is as follows. 1.

Cisco is viewed in IT circles as a thought leader in IOT (we note CEO keynotes on IOT at major trade shows) and has an IOT focused cross business unit product team. Cisco’s IOT interest spans a wide range of use cases – from home automation and digital healthcare to smart manufacturing, smart cities and smart retail.

2.

Cisco has the broadest set of touch points in enterprise and service provider infrastructures – from campus switching, wireless access and security to datacenter servers, switching, core routing and mobile packet core – and a broad portfolio of IT services from campus to the core.

3.

We are constructive on Cisco’s IOT opportunity in 2H15+, given our view that the company has the setup, from a product portfolio, customer base, market share and sales execution POV, to offer a packaged bundle of IOT solutions involving millions of smart objects (i.e. Internet connected wireless sensors) connecting to an Edge Cloud (based on Cisco’s ISR router, etc – for localized switching, routing, security, voice, video and network optimization) and leverage Datacenter Cloud Services for centralized computing, analytics and storage [refer to Figure 1].

Deutsche Bank Securities Inc.

Page 83

6 May 2014 TMT The Internet of Things

4.

The stock could be a potential margin improvement and earnings growth story in CY15+, if the company were to gradually exit from several slower growth and declining margin areas such as lower-end campus switching, parts of SP video and collaboration, optical transport, etc and razor focus on higher margin “software” and “advanced services” opportunities in Private Clouds, Telco Clouds and IOT (refer to our recent CSCO Growth Strategy note.

Our Hold rating on Cisco reflects our near-term caution on the company’s core business (switching and routing in product transition mode) while remaining constructive on CY15+ opportunities in IOT and product refresh cycles in datacenter switching, security, carrier routing, etc.

CommScope The Wireless business segment has been a key driver of CommScope’s overall top line. And while there is nothing specific in Wireless that falls within IOT per se, their DAS solution will be a key enabler of IOT, as it is puts the cellular network closer to the user. We have discussed the possibilities of its DAS solution, both its current one and its ION-e solution, in a recent S2N (S2N #497 – What the DAS market means for CommScope). It is in the Enterprise segment where CommScope has made some interesting acquisitions, mainly within the vein of IOT. Redwood Systems, an intelligent lighting company, and iTRACs, a data center infrastructure management platform, were both acquired when CommScope was held privately. We discussed the potential for iTRACs in our Big Data section within the note. In short, iTRACs has been thought of as the “autoplilot feature on a plane” for data centers and we believe this management tool adds compelling value to data center investments. With Redwood Systems, companies will be able to measure temperature, light and occupancy with these low-voltage, connected LED systems. While the market is likely limited to new construction, we believe that most companies will understand the capex to opex calculus and make the investment. However, it is a competitive market – Philips is a large player in the space – but believe that the market is big enough, and growing quickly enough, to support multiple players. In short, the trends in IOT will impact both Wireless and Enterprise. While Wireless has the general trend of densification working in its favor, it is likely the Enterprise segment that will be impacted by IOT.

F5 Buy-rated F5 is our top midcap idea to play the near-term [CY15/16+] spending on IOT – given our research noting that Network Intelligence and Security are a Phase 1 priority in the multi-year rollout of IOT use cases in enterprise and service provider verticals. A third of the +$2B in incremental networking capex a year (DB view) on IOT use cases could be utilized for Network Intelligence and Security solutions (we note: Layer 4/7 capex is typically a third of datacenter networking capex). This is basis for our view that the Layer 4/7 category leaders (FFIV; BLOX, etc) could see modest TAM expansion in the $100s of Millions a year range from enterprise IOT use cases such as Industrial Cyber Security, Smart Energy,

Page 84

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Smart Retail, etc and also service provider IOT use cases such as e-Healthcare, home automation, etc in the out quarters. Specific IOT touch points and sales opportunities for F5 could be in the areas of: Rollouts of virtual ADCs and HW optimized ADCs in the IOT Datacenter Cloud – for Network Intelligence and Security functions such as: 1) Application and Network Firewalling; 2) Web Traffic Steering; 3) Per-Session IT Policy and Bandwidth Enforcement, 4) Application- and Usage-Specific Billing Metrics, etc. F5’s Application Security and Network Firewalling SW Modules can scale to 100s of Gigabits of packet processing throughput (in comparison, we note that many of F5’s security peers scale to 10s of Gb/s) for mitigating Distributed Denial of Service [DDOS] attacks and sophisticated Web and application layer attacks at corporate servers. More recently, F5 has been offering a Cloud based subscriptions service (leveraging the Versafe acquisition) for detecting and IT reporting of advanced malware threats across IP connected devices – from mobile devices running industry standard OS’ to industry vertical devices running Linux or Windows OS. From a security attach rate POV, it is our view that F5’s Layer 4/7 systems (BIG IP and Viprion) could see the security SW modules attach rate scale from the current +30% of HW platforms sold to around +50%, over the next few years. Internet of Things use cases are a key catalyst for scaling the company’s security solutions sales opportunities, in our view. This is given the significantly higher number of security IT end-points in IOT use cases – which could be the underlying basis for scaling the quantity and types of security modules required to defend against a broad set of attack vectors in IOT Cyber Security use cases. F5’s Access Policy SW Module has an important role to play in controlling end-user access to a wide range of datacenter application servers – at a fine level of granularity – i.e. access control by user name, workgroup affiliation, geographic coordinates, user context, etc. For example: IOT sensors imbedded in various industrial objects can relay security incidence monitoring events to F5’s Access Policy Module – and F5’s TMOS SW is designed to cross-correlate across 1000s of individual security incidence response and access control events – to generate security trend analysis patterns, feedback to IT security policy design, etc. F5’s iRules scripting framework has a core value proposition in IOT use cases – given the relative ease at which simple English language IT and business security policies can be defined and implemented in a matter of minutes – to defend the enterprise architecture against a dynamic set of security threat vectors – which by definition need a dynamic policy creation and implementation framework (offered by iRules). In summary, we believe that the low double-digit FY15 consensus product rev growth expectations factors in mostly high single-digit growth in F5’s core ADC business and few points of incremental growth from the Security and Deutsche Bank Securities Inc.

Page 85

6 May 2014 TMT The Internet of Things

Telco use cases. IOT could be a secular upside driver to consensus product rev growth expectations for F5 – especially as mainstream IOT use cases in Cyber Security, Energy, Retail, Transportation, etc start to scale over the next few years. Network Intelligence and Security are clearly high priority technology spending areas in Internet of Things use cases, in our view.

Ciena Buy-rated Ciena could benefit from sales of higher margin optical Ethernet access products and optical network automation software tools to support IOT Edge Cloud buildouts. We also note opportunities for Ciena’s Carrier Ethernet solution and optical network management software tools in the Smart Energy use case. Undersea cable – carrying optical Ethernet data traffic – is a relevant use case for sales of Ciena’s carrier Ethernet switches at either ends of the offshore and onshore optical WAN links. Ciena’s optical network management SW could be used to dynamically provision optical network transport equipment at either ends of an optical fiber link.

Infoblox Infoblox – a market leader in DNS and IP Address Management [IPAM] solutions – could secularly benefit from the IT spending on IP Address Automation and DNS Security – needed for implementing “Cloud-scale” Internet architectures in the Edge Cloud and Datacenter Cloud – two key themes of our FITT. We have a long view on Infoblox – noting that the company’s IP Address Automation (DNS, DHCP, IP Address Management; collectively IPAM) opportunity could meaningfully scale from – the current +25M enterprise application server IPAM opportunity [refer to our BLOX coverage initiation note for further color] to potentially offering IPAM capabilities for 100s of Millions of Internet Connected Objects in key enterprise and service provider IOT use cases. Infoblox’s GRID architecture – which helps to build a globally correlated view of IP Address Automation – across the extended enterprise, is a key competitive differentiator, in our view, for Infoblox’s design wins in key IOT use cases (versus BLOX’s sub-scale peers). On the DNS Security side, we believe that DNS Security is a need to have (versus a nice to have) in IOT use cases, to safeguard DNS servers situated at the enterprise’s infrastructure borders – i.e. at the network perimeter that connects the enterprise network to the public Internet and to the company’s’ supply chain partners. We note that a vast majority of enterprise DNS servers currently lack adequate security mechanisms to safeguard the DNS servers from security attacks that: Originate from the public Internet (e.g. DDOS attacks targeted at corporate DNS servers); Defend against security attacks that originate from malware-infected employee PCs, smartphones, Tablets, etc, when the employee computing devices are connected to trusted corporate networks. Page 86

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Infoblox recently introduced DNS Security appliances that defend against DNS security attacks that originate from the “outside-in” and the “inside-out” – and these DNS security appliances leverage live DNS reputation feeds to crosscheck all incoming and outgoing DNS queries against a constantly updated live IP address and Web URL feed – so as to identify malicious URLs and hacking attempts on a company’s DNS servers. We estimate Infoblox’s DNS security opportunity for enterprise IT use cases (including IOT use cases) in the mid $100s of million a year range – given our view that DNS security is a next-gen security use case – and the market opportunity for Next-Gen Security solutions in our view is in the +$2B a year range (out of a +$7B a year spend on network security [Infonetics]). We maintain our positive longer-term outlook on Infoblox – the stock trades in the ~2.5-3x EV/Sales range – a discount to its small cap tech peer group [basis: FY15 consensus view].

Juniper We are cautious on Hold-rated Juniper in IOT networking buildouts. Juniper, in our view, lacks a competitive best-in-class Enterprise Edge Cloud and Datacenter Leaf and Spine Cloud switching portfolio. Juniper is likely to play into a narrower set of opportunities in IOT, i.e. primarily in out-year telco IOT use cases that require Juniper’s Contrail SW for service creation on demand. For example: a telco could use the Contrail SW for activating a series of Layer 2/7 virtual networking features within its datacenter for automatically setting up a home automation service for an end-user who signs up for the service via the carrier’s IOT web portal. We see a limited role for Juniper in the access segments of IOT use cases, i.e. in WiFi, campus switching or in campus and branch security. Our cautious view is based on Juniper’s low single-digit market share in enterprise networking and lack of meaningful competitive advantage versus market leader Cisco across a wide range of IOT enterprise networking segments.

Riverbed Hold-rated Riverbed could modestly benefit from enterprise IOT use cases in CY15/16+, primarily in network performance monitoring by leveraging Riverbed’s Cascade product portfolio. That said, we remain cautious on Riverbed, noting +70% of Riverbed’s FY revs are exposed to a declining growth WAN optimization (Steelhead) portfolio.

Cavium While current levels suggest a neutral risk/reward equation – basis for our Hold rating – given, stock trading at ~25x P/E for ~20% consensus CY15 rev growth expectation, we have a constructive longer-term view on Cavium’s new multicore processors: specifically, Octeon3, Liquid IO and Neuron TCAM chips – which are relevant for IOT enterprise and service provider use cases. In the Datacenter Cloud, Cavium’s Liquid IO application acceleration processors could benefit from IOT use cases such as Big Data Analytics, Deutsche Bank Securities Inc.

Page 87

6 May 2014 TMT The Internet of Things

processing high volumes of aggregated and encrypted IOT sensor data (SSL Offload, data compression, etc) – as HW acceleration add-ons in datacenter servers, etc. Similarly, Cavium’s Nitrox processors fit into HW acceleration use cases in Layer 4/7 ADC and Next-Gen Security appliances. The Neuron TCAM chip could see potential design wins in high-end Terabitscale routers such as Cisco’s NCS, etc – which are likely to be used in service provider IOT use cases such as for handling millions of IOT transactional events in home automation, e-Healthcare, Smart Vehicles, etc. For the present, we remain with our Hold rating on the stock – given the rich trading multiple – and as we gain further field color on Cavium’s market opportunity in mainstream IOT Datacenter Cloud use cases.

Ubiquiti Uniquiti is able to proliferate a number of markets quickly, given its low cost and network of distributors around the world. Most investors know the company from its airMax and Unifi product lines; however, it has three other platforms – airVision, mFi and EdgeMax. airVision was released a few years ago and was riddled with issues that eventually tempered demand for the product. EdgeMax is new, but is purpose built for WISP use in its network operating centers and access points and is not necessarily an IOT product. mFi, on the other hand, is Ubqiuiti’s foray into machine to machine communication. The mFi platform includes everything from smart power strips to sensors (motion, light, temperature and doors). While we believe the devices have meaningful utility, we believe their channel and its end-customer (as it stands today) will likely not create the demand necessary to make this platform a meaningful contributor to the top line. If instead, these devices were easy enough for a consumer to install, and available in known retail locations, our sense is that the traction would be much more significant. As it stands, we see it as more of a hobby of the WISP operator, or other tech savvy people.

While we believe the devices have meaningful utility, we believe their channel, and its end-customer (as it stands today) will likely not create the demand necessary to make this platform a meaningful contributor to the top line

Aruba and Ruckus We grouped Ruckus and Aruba together here as there is not necessarily any one product or platform directed at IOT with either company; however, both should see a general uplift from the broad trend of connectivity, as well as other trends which go along with this such as: 802.11ac adoption, location analytics, BYOD, Hotspot 2.0 and cloud managed WLAN services. All of this will play a role in bringing in faster wireless connectivity closer to the user. Additionally, some IOT modules will connect via WiFi and will require a sort of dynamic usage of the spectrum – it will be up to the AP to understand this. This is where Aruba and Ruckus could differentiate their wares, as both have distanced themselves so far by incrementally adding intelligence into their AP’s and going further to integrate IOT elements would not be a reach for either company. In short, we believe that these two will eventually be beneficiaries of IOT, but it will likely not happen in the near term.

Page 88

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Audience Audience is known for its audio and voice processors. The company has Samsung as a significant customer, as well as Chinese vendors like Lenovo, ZTE and Xiaomi. The audio and voice processing unit will likely continue to move alongside the trajectory of its OEM handset customers. But where Audience could make an inroad into IOT is with its Voice Q and Motion Q technologies, which wakes and sleeps a device based on voice and motion recognition, thereby efficiently managing battery life. The company recently demoed versions of the processor at the Consumer Electronics Show and at Mobile World Congress. It will be integrated into some of its eS700 series of processors, which will ship in the second half of this year. While it will likely gain initial traction in the smartphone realm, we could see this being used in remote controls, wearables, fitness and health devices, mobile computers and other battery powered devices. Going back to our original thesis, while we believe that it will be some time before this type of processing capability sees significant growth, we do think that it will be eventually included in adjacent markets and Audience could see meaningful upside as a result of this – from our viewpoint, though, this would be a number years away.

Ixia Ixia could see a modest set of opportunities in network performance visibility monitoring in IOT use cases given our view that SDN controllers from Cisco, etc could be cost-competitive alternatives for SW based network visibility monitoring versus their HW based counterparts over the next few years.

Companies outside of our coverage area 2lemetry: 2lementary provides data management and analytics solutions for IOT environments. It provides the middleware, development layer and cloud services so that data generated in IOT can be managed and analyzed. Affinegy: Affinegy manufactures products that provide network management solutions for IOT, specifically as it relates to the home environment. Its network management solutions are focused to support cloud based services in IOT environments. AT&T: Most recognize ATT as one of the leading carriers in North America. As it relates to IOT though, the company provides IP-based communication services to enterprises, an expanding business line for the company. Additionally, it owns the largest Wi-Fi network in United States. Beechwoods Software, Inc: Beechwoods creates solutions for the embedded software market. As this relates to IOT, it helps companies implement ideas on specific devices. The company got its start by helping DirectTV develop its settop satellite box. Belkin: Belkin has an entire division called WeMo devoted to IOT, with connected lighting and switches. While mostly known for its wireless routers and computing peripherals, Belkin likely sees home automation as a key growth area.

Deutsche Bank Securities Inc.

Page 89

6 May 2014 TMT The Internet of Things

CA Engineering: CA Engineering provides solutions related to storm water management, wastewater design, water supply design and water resources engineering. Controlled Network Solutions (CNS): CNS has been providing hosting and IT support to its customers that are operating in various industries such as finance, healthcare and manufacturing. D-Link: Similar to TP-Link, D-Link is a manufacturer of connectivity products and often is able to design and manufacture them at compelling price points. DoubleTwist: DoubleTwist developed a platform to deliver music in a device agnostic manner. The company is writing to the AllJoyn protocol and is likely going to be one of the earlier beneficiaries of the IOT idea. Fon: Fon is WiFi product manufacturer and global WiFi network. The company leverages the extra capacity of subscribers’ home wireless routers in order to provide service for other Fon members, in the area. To date, Fon has wireless capacity in 100 countries and over 7 million hotspots. Haier: Founded in 1984 in China, Haier is one of the largest producers of white goods in the world. It has 80,000 employees, five R&D centers and operates in more than 100 countries. The company has a number of efforts to embed its products with IOT capabilities; it was one of the first to sign up for Qualcomm’s AllJoyn network. Harman: Harman International Industries includes brands JBL, Mark Levinson, Lexicon, Logic 7, Becker and Infinity. The company manufactures electronic systems and audio products, as well as develops mobile applications. The company will likely be a player in the early pocket of IOT investment – the connected car. HTC: HTC is a handset manufacturer which, given its reliance on Android OS, has made strides in terms of IOT connectivity. The company has joined and is an active member of the IOT AllSeen alliance. iControl Networks: a software based home control platform for managing the connected home. Applications include music, energy monitoring, lighting control, traffic, weather and news alerts. Imagination: The company provides various electronic goods manufacturers with its System-on-Chip intellectual property (SoC IP). Its SoC IP powers some of the more advanced electronic goods. For further information on this company please contact Johannes Schaller. Impinj: Impinj is a major provider of ultra high frequency RFiD solutions, which are used to locate, identify and authenticate items for applications such as retail inventory management, warehousing and logistics. The company was founded in 2000 and has raised over $110 million in venture funding. Jawbone: Jawbone has developed wearables and audio devices for over 10 years. The company has done well by designing products that appear sleek but are also technologically innovative. The company’s products include fitness wearables, wireless speakers and headsets. Most of its products operate via

Page 90

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Bluetooth technology. The company has built a record of being first to market but with products that are best in class. Kii Corporation: Kii was formed by the merger of Synclore Corporation and Servo Software in July 2010. The company created the Kii Platform, which helps app developers by providing cloud technology solutions and other development solutions. LIFX: LIFX was Kickstarter project that raised $1.3 million in three days. It produced an LED light bulb, equipped with Wi-Fi, which can be controlled by use of its application. LG: LG is one of the largest electronic goods producers in the world. Based in Seoul, South Korea, its main product verticals are home appliances, entertainment, energy solutions, air conditioning and semiconductors. The company supported AllJoyn protocols in its most recent iteration of smart TV releases and released a smart light bulb within the past two weeks. Liteon: Liteon, a Taiwanese company, is a leading producer of optoelectronic devices. The company develops products like photo-interrupters and photocouplers. M87: M87 developed Routing at the Edge (RATE), which is a software solution designed to move the intelligence to the device rather than the network, thereby allowing the device to leverage the best connection, be it the macro cell or another handset. The company is located in Austin Texas and has raised $3 million in funding. Metaio: Metaio is one of the leaders in augmented reality systems. We first met the company at Mobile World Congress in Barcelona and were impressed with their demos. Augmented reality has a growing number of use cases, currently providing solutions like the placement of virtual Ikea furniture in the house to understanding the elements of a car during repair. Over time, the technology will increasingly leverage wearable technology, becoming more energy efficient through the use of its cloud engine. All of this will help augmented reality to become a more seamless experience and likely more ubiquitous with time. Moxtreme: Moxtreme designs products to help service providers and OEMS manufacture products with smart connecting ability. Musaic: Musaic is a London based company that produces wireless HiFi systems centered on the ability to play music from Android devices, Apple devices, Macs, PCs and other sources. Muzzley: Muzzley provides a middleware in the cloud by the help of which OEM manufacturers and service providers can create their own custom user interface to control the connected devices in an IOT environment. Panasonic: Panasonic has over 634 companies under its corporate umbrella, and with that, designs and manufactures a broad range of electronic goods under the Panasonic brand. The company has developed M2M and IOT devices and chipsets, most recently releasing a multi-mode M2M chipset that would allow devices to communicate autonomously over various wireless protocols. Deutsche Bank Securities Inc.

Page 91

6 May 2014 TMT The Internet of Things

Patavina Technologies: Founded in 2009, as a result of spin off of University of Padova, Patavina Technologies provides system and software design services to its customers. Its main product, ThinkIP, is a device controlling solution for IOT. Quirky: A different company, Quirky crowd sources ideas and then designs and manufactures the best ones. It designs products across a number of verticals (not necessarily all related to IOT) and has developed sensors and power strips that have connectivity and leverage the IOT idea. R/GA: R/GA is an accelerator with a sole focus on developing start-up companies in the connected device space. The company has a number of large clients (e.g. Google; Unilever) for which it helps develop connected devices. Sears: Sears is one of the largest retailers in Canada and United States with more than 3,900 stores, including specialty and full line retail stores. In some product areas, the company manufactures its own products and in doing so, has made inroads into IOT. Silicon Image: Silicon Image provides connectivity technology and semiconductor products for HD playback over consumer devices like TV, tablets and smartphones. The company played a role with the establishment of DVI and HDMI standards in the digital industry. Silicon Labs: Silicon Labs is a semiconductor company, which designs chipsets for sensors, portable devices and gateways. The company has an extensive history of creating low-power processors for elements, which we would now place into the IOT realm. Skyworks: Skyworks is one company that is seeing early exposure to the IOT, resulting in their Broad Mix business segment (infrastructure; connected home; m2m; medical; auto) to become a main driver for overall growth despite it only being 30% of the total business. The company had its solutions incorporated into a number of products including TV’s, wireless lighting, auto navigation assistance, intelligent thermostats and wearable medical solutions. The company has front-end modules for WiFi 802.11ac, switches for remote car entry, converters, touch screen drivers, amplifiers and other, unlicensed front end modules (e.g. Zigbee for remotes and 900 Mhz for metering). All told, the company’s analogue processors are seeing meaningful traction with IOT given their price relative to performance. Smartthings: With $15 million in funding, Smartthings designs and manufactures a number of physical sensors, along with a gateway, for home automation. The sensors include those for doors, locks, motion, lights, damages and dangers and the hub controls all elements. Put together, they allow a person to easily automate a number of home elements and do so in an easy-to-use manner. Sproutling: Sproutline is a startup THAT develops wearable monitors for babies. It monitors a baby’s movement, heart rate, temperature, room’s humidity level, light level and noise level. The Sprosty Network: Sprosty provides technological support to firms engaged in development of wireless solutions, emerging technologies and consumer technology. Page 92

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Technicolor: Technicolor manufactures DVD and CD and provides content and media services like visual effects, sound services, game design and animation. In terms of IOT, the company has developed multi-screen, OTT, smart home and ultra HD solutions. TP Link: TP Link is best known for its inexpensive WLAN networking products, which include switches, routers, ADSL, powerline adapters, networking adapters and media converters. The company will likely integrate the consensus protocol within its devices, and given its cost competiveness, will be one of the main volume producers of IOT connected gateways. Trimble: Trimble has long created devices that leverage wireless technology, be it cellular, GPS, or WLAN, which are more effective and easier to use than their older, analogue counterparts. The company makes connected devices for industries such as agriculture, construction, transportation and geospatial. Tuxera: Tuxera’s NTFS and exFAT solutions are designed for the file systems space. Tuxera provides low power, high performance mobile and embedded solutions. In 2012, BMW and Audi became the first cars to have Tuxera file systems installed in them. Two Bulls: Two Bulls develops mobile & web apps, augmented reality and enterprise solutions. It is one of the preferred developers of Qualcomm for augmented reality and is also a member of Apple Consultants Network. Vestel: Founded in 1984, Vestal is a Turkey based company known for manufacturing a wide range of consumer electronics. It is also engaged in the development of software and technology related to consumer electronics and multimedia communication. Weaved: Weaved provides cloud services, mobile apps, software and technology for devices in the IOT environment, essentially making it easier for a device to become an IOT device. Weaved’s goal is to therefore quicken the time-to-market for new products. Wilocity – Founded in 2007, Wilocity is based in Israel. It is famous for its low power 60 GHZ wireless chipset with data rates up to 7 Gbps. It provides wireless connectivity and networking solutions with technological partners, including Marvell Semiconductors and Qualcomm. Winner Micro: Based in China, Winner Micro develops IOT centric wireless communication chips and solutions, and does so for a variety of verticals including home, medical and industrial fields Withings: The company is focused solely on the medical devices market. To date, the company has raised $34 million. Products range from a smart baby monitor to a smart body analyzer. The company uses smartphones and computers as bridges to the network and to date, have developed a network of over 100 developer and device partners.

Deutsche Bank Securities Inc.

Page 93

6 May 2014 TMT The Internet of Things

Appendix 1 Important Disclosures Additional information available upon request Disclosure checklist Company

Ticker

Recent price*

Disclosure

Commscope

COMM.OQ

27.84 (USD) 2 May 14

1,2,6,7,8

Cisco Systems

CSCO.OQ

22.94 (USD) 2 May 14

1,2,6,7,8,14,15

F5 Networks

FFIV.OQ

103.72 (USD) 2 May 14

2

Qualcomm

QCOM.OQ

78.99 (USD) 2 May 14

2,6,7,8,14,15

*Prices are sourced from local exchanges via Reuters, Bloomberg and other vendors. Data is sourced from Deutsche Bank and subject companies

Important Disclosures Required by U.S. Regulators Disclosures marked with an asterisk may also be required by at least one jurisdiction in addition to the United States. See Important Disclosures Required by Non-US Regulators and Explanatory Notes. 1.

Within the past year, Deutsche Bank and/or its affiliate(s) has managed or co-managed a public or private offering for this company, for which it received fees.

2.

Deutsche Bank and/or its affiliate(s) makes a market in securities issued by this company.

6.

Deutsche Bank and/or its affiliate(s) owns one percent or more of any class of common equity securities of this company calculated under computational methods required by US law.

7.

Deutsche Bank and/or its affiliate(s) has received compensation from this company for the provision of investment banking or financial advisory services within the past year.

8.

Deutsche Bank and/or its affiliate(s) expects to receive, or intends to seek, compensation for investment banking services from this company in the next three months.

14. Deutsche Bank and/or its affiliate(s) has received non-investment banking related compensation from this company within the past year. 15. This company has been a client of Deutsche Bank Securities Inc. within the past year, during which time it received non-investment banking securities-related services.

Important Disclosures Required by Non-U.S. Regulators Please also refer to disclosures in the Important Disclosures Required by US Regulators and the Explanatory Notes. 1.

Within the past year, Deutsche Bank and/or its affiliate(s) has managed or co-managed a public or private offering for this company, for which it received fees.

2.

Deutsche Bank and/or its affiliate(s) makes a market in securities issued by this company.

6.

Deutsche Bank and/or its affiliate(s) owns one percent or more of any class of common equity securities of this company calculated under computational methods required by US law.

7.

Deutsche Bank and/or its affiliate(s) has received compensation from this company for the provision of investment banking or financial advisory services within the past year.

For disclosures pertaining to recommendations or estimates made on securities other than the primary subject of this research, please see the most recently published company report or visit our global disclosure look-up page on our website at http://gm.db.com/ger/disclosure/DisclosureDirectory.eqsr

Page 94

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Analyst Certification The views expressed in this report accurately reflect the personal views of the undersigned lead analyst about the subject issuers and the securities of those issuers. In addition, the undersigned lead analyst has not and will not receive any compensation for providing a specific recommendation or view in this report. Brian Modoff Historical recommendations and target price: Commscope (COMM.OQ) (as of 5/2/2014) 30.00

Previous Recommendations 3

25.00

2

20.00

Security Price

4

1

Strong Buy Buy Market Perform Underperform Not Rated Suspended Rating Current Recommendations

15.00

Buy Hold Sell Not Rated Suspended Rating

10.00

*New Recommendation Structure as of September 9,2002

5.00

0.00 Oct 13

Jan 14

Apr 14

Date

1.

12/04/2013:

Upgrade to Buy, Target Price Change USD22.00

3.

04/17/2014:

Buy, Target Price Change USD28.00

2.

02/21/2014:

Buy, Target Price Change USD24.00

4.

04/30/2014:

Buy, Target Price Change USD30.00

Deutsche Bank Securities Inc.

Page 95

6 May 2014 TMT The Internet of Things

Historical recommendations and target price: Cisco Systems (CSCO.OQ) (as of 5/2/2014) 30.00

Previous Recommendations 12 10 11

25.00

4

Security Price

1

2 3

14

9

5 20.00

13

6

78

Strong Buy Buy Market Perform Underperform Not Rated Suspended Rating Current Recommendations

15.00

Buy Hold Sell Not Rated Suspended Rating

10.00

*New Recommendation Structure as of September 9,2002

5.00

0.00 May 11 Aug 11 Nov 11 Feb 12 May 12 Aug 12 Nov 12 Feb 13 May 13 Aug 13 Nov 13 Feb 14

Date

1.

05/12/2011:

Hold, Target Price Change USD19.00

8.

11/14/2012:

2.

07/11/2011:

Hold, Target Price Change USD18.00

9.

02/04/2013:

Buy, Target Price Change USD22.00 Buy, Target Price Change USD24.00

3.

08/11/2011:

Hold, Target Price Change USD17.00

10. 05/16/2013:

Buy, Target Price Change USD25.00

4.

11/10/2011:

Upgrade to Buy, Target Price Change USD22.00

11. 07/01/2013:

Buy, Target Price Change USD26.00

5.

02/09/2012:

Buy, Target Price Change USD24.00

12. 08/02/2013:

Buy, Target Price Change USD28.00

6.

05/10/2012:

Buy, Target Price Change USD22.00

13. 11/14/2013:

Downgrade to Hold, Target Price Change USD25.00

7.

11/04/2012:

Buy, Target Price Change USD21.00

14. 02/13/2014:

Hold, Target Price Change USD24.00

Page 96

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Historical recommendations and target price: F5 Networks (FFIV.OQ) (as of 5/2/2014) 160.00

Previous Recommendations 6

140.00

7 120.00

4

8

16

5

Security Price

9 100.00 80.00

3

10

15 11 13

12

14

17

Strong Buy Buy Market Perform Underperform Not Rated Suspended Rating Current Recommendations

12

Buy Hold Sell Not Rated Suspended Rating

60.00 40.00

*New Recommendation Structure as of September 9,2002

20.00 0.00 May 11 Aug 11 Nov 11 Feb 12 May 12 Aug 12 Nov 12 Feb 13 May 13 Aug 13 Nov 13 Feb 14

Date

1.

08/12/2011:

Buy, Target Price Change USD110.00

10. 10/25/2012:

2.

09/04/2011:

Buy, Target Price Change USD95.00

11. 04/05/2013:

Buy, Target Price Change USD95.00

3.

10/26/2011:

Buy, Target Price Change USD110.00

12. 04/25/2013:

Buy, Target Price Change USD93.00

4.

11/13/2011:

Buy, Target Price Change USD125.00

13. 07/25/2013:

Buy, Target Price Change USD100.00

5.

01/19/2012:

Buy, Target Price Change USD130.00

14. 10/24/2013:

Buy, Target Price Change USD107.00

6.

04/02/2012:

Buy, Target Price Change USD135.00

15. 01/23/2014:

Buy, Target Price Change USD120.00

7.

04/19/2012:

Buy, Target Price Change USD150.00

16. 03/03/2014:

Buy, Target Price Change USD130.00

8.

05/24/2012:

Buy, Target Price Change USD140.00

17. 04/24/2014:

Buy, Target Price Change USD135.00

9.

07/19/2012:

Buy, Target Price Change USD136.00

Deutsche Bank Securities Inc.

Buy, Target Price Change USD116.00

Page 97

6 May 2014 TMT The Internet of Things

Historical recommendations and target price: Qualcomm (QCOM.OQ) (as of 5/2/2014) 90.00

Previous Recommendations 11

80.00

5 70.00

Security Price

60.00

9

8 4

1 2

6

Strong Buy Buy Market Perform Underperform Not Rated Suspended Rating

10

7

3

50.00

Current Recommendations Buy Hold Sell Not Rated Suspended Rating

40.00 30.00 20.00

*New Recommendation Structure as of September 9,2002

10.00 0.00 May 11 Aug 11 Nov 11 Feb 12 May 12 Aug 12 Nov 12 Feb 13 May 13 Aug 13 Nov 13 Feb 14

Date

1.

07/21/2011:

Buy, Target Price Change USD67.00

7.

11/08/2012:

Buy, Target Price Change USD74.00

2.

09/04/2011:

Buy, Target Price Change USD65.00

8.

01/31/2013:

Buy, Target Price Change USD78.00

3.

11/03/2011:

Buy, Target Price Change USD70.00

9.

08/17/2013:

Buy, Target Price Change USD80.00

4.

02/02/2012:

Buy, Target Price Change USD72.00

10. 11/07/2013:

Buy, Target Price Change USD78.00

5.

03/28/2012:

Buy, Target Price Change USD75.00

11. 03/21/2014:

Buy, Target Price Change USD86.00

6.

07/19/2012:

Buy, Target Price Change USD70.00

Equity rating key Buy: Based on a current 12- month view of total share-holder return (TSR = percentage change in share price from current price to projected target price plus pro-jected dividend yield ) , we recommend that investors buy the stock. Sell: Based on a current 12-month view of total shareholder return, we recommend that investors sell the stock Hold: We take a neutral view on the stock 12-months out and, based on this time horizon, do not recommend either a Buy or Sell. Notes: 1. Newly issued research recommendations and target prices always supersede previously published research. 2. Ratings definitions prior to 27 January, 2007 were:

Equity rating dispersion and banking relationships 500 450 400 350 300 250 200 150 100 50 0

49 %

49 %

56 % 42 % 2 % 32 % Buy

Hold

Companies Covered

Sell

Cos. w/ Banking Relationship

North American Universe

Buy: Expected total return (including dividends) of 10% or more over a 12-month period Hold: Expected total return (including dividends) between -10% and 10% over a 12month period Sell: Expected total return (including dividends) of -10% or worse over a 12-month period

Page 98

Deutsche Bank Securities Inc.

6 May 2014 TMT The Internet of Things

Regulatory Disclosures 1. Important Additional Conflict Disclosures Aside from within this report, important conflict disclosures can also be found at https://gm.db.com/equities under the "Disclosures Lookup" and "Legal" tabs. Investors are strongly encouraged to review this information before investing.

2. Short-Term Trade Ideas Deutsche Bank equity research analysts sometimes have shorter-term trade ideas (known as SOLAR ideas) that are consistent or inconsistent with Deutsche Bank's existing longer term ratings. These trade ideas can be found at the SOLAR link at http://gm.db.com.

3. Country-Specific Disclosures Australia and New Zealand: This research, and any access to it, is intended only for "wholesale clients" within the meaning of the Australian Corporations Act and New Zealand Financial Advisors Act respectively. Brazil: The views expressed above accurately reflect personal views of the authors about the subject company(ies) and its(their) securities, including in relation to Deutsche Bank. The compensation of the equity research analyst(s) is indirectly affected by revenues deriving from the business and financial transactions of Deutsche Bank. In cases where at least one Brazil based analyst (identified by a phone number starting with +55 country code) has taken part in the preparation of this research report, the Brazil based analyst whose name appears first assumes primary responsibility for its content from a Brazilian regulatory perspective and for its compliance with CVM Instruction # 483. EU countries: Disclosures relating to our obligations under MiFiD can be found at http://www.globalmarkets.db.com/riskdisclosures. Japan: Disclosures under the Financial Instruments and Exchange Law: Company name - Deutsche Securities Inc. Registration number - Registered as a financial instruments dealer by the Head of the Kanto Local Finance Bureau (Kinsho) No. 117. Member of associations: JSDA, Type II Financial Instruments Firms Association, The Financial Futures Association of Japan, Japan Investment Advisers Association. Commissions and risks involved in stock transactions - for stock transactions, we charge stock commissions and consumption tax by multiplying the transaction amount by the commission rate agreed with each customer. Stock transactions can lead to losses as a result of share price fluctuations and other factors. Transactions in foreign stocks can lead to additional losses stemming from foreign exchange fluctuations. "Moody's", "Standard & Poor's", and "Fitch" mentioned in this report are not registered credit rating agencies in Japan unless Japan or "Nippon" is specifically designated in the name of the entity. Reports on Japanese listed companies not written by analysts of Deutsche Securities Inc. (DSI) are written by Deutsche Bank Group's analysts with the coverage companies specified by DSI. Qatar: Deutsche Bank AG in the Qatar Financial Centre (registered no. 00032) is regulated by the Qatar Financial Centre Regulatory Authority. Deutsche Bank AG - QFC Branch may only undertake the financial services activities that fall within the scope of its existing QFCRA license. Principal place of business in the QFC: Qatar Financial Centre, Tower, West Bay, Level 5, PO Box 14928, Doha, Qatar. This information has been distributed by Deutsche Bank AG. Related financial products or services are only available to Business Customers, as defined by the Qatar Financial Centre Regulatory Authority. Russia: This information, interpretation and opinions submitted herein are not in the context of, and do not constitute, any appraisal or evaluation activity requiring a license in the Russian Federation. Kingdom of Saudi Arabia: Deutsche Securities Saudi Arabia LLC Company, (registered no. 07073-37) is regulated by the Capital Market Authority. Deutsche Securities Saudi Arabia may only undertake the financial services activities that fall within the scope of its existing CMA license. Principal place of business in Saudi Arabia: King Fahad Road, Al Olaya District, P.O. Box 301809, Faisaliah Tower - 17th Floor, 11372 Riyadh, Saudi Arabia. United Arab Emirates: Deutsche Bank AG in the Dubai International Financial Centre (registered no. 00045) is regulated by the Dubai Financial Services Authority. Deutsche Bank AG - DIFC Branch may only undertake the financial services activities that fall within the scope of its existing DFSA license. Principal place of business in the DIFC: Dubai International Financial Centre, The Gate Village, Building 5, PO Box 504902, Dubai, U.A.E. This information has been distributed by Deutsche Bank AG. Related financial products or services are only available to Professional Clients, as defined by the Dubai Financial Services Authority.

Deutsche Bank Securities Inc.

Page 99

David Folkerts-Landau Group Chief Economist Member of the Group Executive Committee Guy Ashton Global Chief Operating Officer Research Michael Spencer Regional Head Asia Pacific Research

Marcel Cassard Global Head FICC Research & Global Macro Economics

Ralf Hoffmann Regional Head Deutsche Bank Research, Germany

Richard Smith and Steve Pollard Co-Global Heads Equity Research

Andreas Neubauer Regional Head Equity Research, Germany

Steve Pollard Regional Head Americas Research

International locations Deutsche Bank AG Deutsche Bank Place Level 16 Corner of Hunter & Phillip Streets Sydney, NSW 2000 Australia Tel: (61) 2 8258 1234

Deutsche Bank AG Große Gallusstraße 10-14 60272 Frankfurt am Main Germany Tel: (49) 69 910 00

Deutsche Bank AG London 1 Great Winchester Street London EC2N 2EQ United Kingdom Tel: (44) 20 7545 8000

Deutsche Bank Securities Inc. 60 Wall Street New York, NY 10005 United States of America Tel: (1) 212 250 2500

Deutsche Bank AG Filiale Hongkong International Commerce Centre, 1 Austin Road West,Kowloon, Hong Kong Tel: (852) 2203 8888

Deutsche Securities Inc. 2-11-1 Nagatacho Sanno Park Tower Chiyoda-ku, Tokyo 100-6171 Japan Tel: (81) 3 5156 6770

Global Disclaimer The information and opinions in this report were prepared by Deutsche Bank AG or one of its affiliates (collectively "Deutsche Bank"). The information herein is believed to be reliable and has been obtained from public sources believed to be reliable. Deutsche Bank makes no representation as to the accuracy or completeness of such information. Deutsche Bank may engage in securities transactions, on a proprietary basis or otherwise, in a manner inconsistent with the view taken in this research report. In addition, others within Deutsche Bank, including strategists and sales staff, may take a view that is inconsistent with that taken in this research report. Opinions, estimates and projections in this report constitute the current judgement of the author as of the date of this report. They do not necessarily reflect the opinions of Deutsche Bank and are subject to change without notice. Deutsche Bank has no obligation to update, modify or amend this report or to otherwise notify a recipient thereof in the event that any opinion, forecast or estimate set forth herein, changes or subsequently becomes inaccurate. Prices and availability of financial instruments are subject to change without notice. This report is provided for informational purposes only. It is not an offer or a solicitation of an offer to buy or sell any financial instruments or to participate in any particular trading strategy. Target prices are inherently imprecise and a product of the analyst judgement. In August 2009, Deutsche Bank instituted a new policy whereby analysts may choose not to set or maintain a target price of certain issuers under coverage with a Hold rating. In particular, this will typically occur for "Hold" rated stocks having a market cap smaller than most other companies in its sector or region. We believe that such policy will allow us to make best use of our resources. Please visit our website at http://gm.db.com to determine the target price of any stock. The financial instruments discussed in this report may not be suitable for all investors and investors must make their own informed investment decisions. Stock transactions can lead to losses as a result of price fluctuations and other factors. If a financial instrument is denominated in a currency other than an investor's currency, a change in exchange rates may adversely affect the investment. Past performance is not necessarily indicative of future results. Deutsche Bank may with respect to securities covered by this report, sell to or buy from customers on a principal basis, and consider this report in deciding to trade on a proprietary basis. Unless governing law provides otherwise, all transactions should be executed through the Deutsche Bank entity in the investor's home jurisdiction. In the U.S. this report is approved and/or distributed by Deutsche Bank Securities Inc., a member of the NYSE, the NASD, NFA and SIPC. In Germany this report is approved and/or communicated by Deutsche Bank AG Frankfurt authorized by the BaFin. In the United Kingdom this report is approved and/or communicated by Deutsche Bank AG London, a member of the London Stock Exchange and regulated by the Financial Conduct Authority for the conduct of investment business in the UK and authorized by the BaFin. This report is distributed in Hong Kong by Deutsche Bank AG, Hong Kong Branch, in Korea by Deutsche Securities Korea Co. This report is distributed in Singapore by Deutsche Bank AG, Singapore Branch or Deutsche Securities Asia Limited, Singapore Branch (One Raffles Quay #18-00 South Tower Singapore 048583, +65 6423 8001), and recipients in Singapore of this report are to contact Deutsche Bank AG, Singapore Branch or Deutsche Securities Asia Limited, Singapore Branch in respect of any matters arising from, or in connection with, this report. Where this report is issued or promulgated in Singapore to a person who is not an accredited investor, expert investor or institutional investor (as defined in the applicable Singapore laws and regulations), Deutsche Bank AG, Singapore Branch or Deutsche Securities Asia Limited, Singapore Branch accepts legal responsibility to such person for the contents of this report. In Japan this report is approved and/or distributed by Deutsche Securities Inc. The information contained in this report does not constitute the provision of investment advice. In Australia, retail clients should obtain a copy of a Product Disclosure Statement (PDS) relating to any financial product referred to in this report and consider the PDS before making any decision about whether to acquire the product. Deutsche Bank AG Johannesburg is incorporated in the Federal Republic of Germany (Branch Register Number in South Africa: 1998/003298/10). Additional information relative to securities, other financial products or issuers discussed in this report is available upon request. This report may not be reproduced, distributed or published by any person for any purpose without Deutsche Bank's prior written consent. Please cite source when quoting. Copyright © 2014 Deutsche Bank AG

GRCM2014PROD031966