The Internet of Things is a moving target that defies tidy description. Essentially a technology-enabled solutions approach to solving problems, IoT is comprised of multiple moving parts that when deployed together can offer new, unimagined levels of service delivery for businesses, citizens and the planet. Popularization of the term – and its application to a broad range of use cases – has helped to stimulate new thinking around technology solutions; however, promotion of IoT has also served to distort understanding through oversimplification. The recent Wavefront Summit, an annual event hosted by the Vancouver-based Wavefront Centre of Excellence in Ottawa this year, proved an exception to this rule, presenting a diversity of views that captured a good sense of the complexity that this emerging area represents.
Wavefront focus at the Summit on IoT is a logical outcome of synergies and common sense: while the CoE is devoted to accelerating the growth of Canadian wireless innovation, networking communications, along with data analytics and sensor technologies, are a foundational component of IoT implementations. At another level altogether, interest in IoT is high. As Brad Lowe, Wavefront VP of national accelerator operations and Summit emcee noted, attendance at this year’s event hit an all-time high of 400, with CEOs and company founders making up 30 percent of the total, VPs 10 percent, and public sector representatives 12 percent of the presentation audience and event’s B2B meeting groups. What’s driving this interest? Wavefront CEO James Maynard’s answer at the Summit was economic potential: citing Cisco and MIT research, Maynard positioned Canadian IoT economic opportunity at four times that which would be available through trade with China.
Imagining the possibilities
Track two at the Summit launched with a visioning exercise, a practice that Cisco is well versed in and which Kim Devooght, VP, public sector, Cisco Canada, was able to support with multiple examples. Devooght shared Cisco’s own experience transforming its workspace over a six month period with collaboration tools aimed at reducing travel and consumerization technologies designed to address demographic shifts that put next generation workers and their mobile IT habits at the forefront. He also noted the role of government in creating policies and programs that can provide the foundation for IoT implementation, as well as their manifestation – the ‘Digital France’ initiative, for example, or smart cities like Stockholm and Barcelona. Devooght used Dundee Precious Metals, which deployed an IP-based wireless network to connect underground phones, tablets, surveillance cameras, Telepresence systems, as well as controllers in the mill, building lights, fans, power and the blasting system, as an industry demonstration of the power of IoT. Targeting a 30 percent increase in production at its flagship implementation in Bulgaria, Dundee achieved a 400 percent improvement through its IoT deployment.
Interestingly, from a public sector perspective, Devooght argued that value in IoT may be derived from technology, but even more importantly from new business models: approximately 70 percent of the value you can capture, he explained, is agency specific, while 30 percent is generated from outside the organization, a notion that may have appeal for governments tasked with stimulating growth in other sectors. According to Devooght, to identify and capture this value, the public sector must abandon “the five year plan” in favour of “fast innovation” with a proven methodology for “interconnection in IoE” which can produce value through: an increase in employee productivity, the connection of militarized defense and intelligence, cost reduction, an improved citizen experience, and additional revenue derived from new opportunities and increased compliance. To support uptake of IoT, he concluded his presentation with advice for governments at the policy and agency levels: governments must create the right business climate and regulatory environment to encourage investment in Canada, while ensuring there is appropriate infrastructure investment; individual projects proceed by determining the level of interest in IoT within an organization, harnessing complementary insights on IoT and identifying specific opportunities, and by learning from others while building a local support group.
IoT mechanics – the data analytics component
Paul Terry, CEO of data warehouse provider PHEMI, also engaged in some visioning in his presentation “Big Data, the Internet of Things and Analytics,” – noting, for example, US President Obama’s ‘precision medicine’ proposal for sequencing the DNA of a million people to support the development of personalized medicine – but his real focus was on the data enablement of IoT. According to Terry, “Big Data is a shoe box” or “a small supercomputer” where technologies like storage and compute are becoming increasingly commoditized, and there are now ingest rates of approximately of 100 million transactions a second. Price drops – to 6 cents per gigaflop for storage – mean that storage and compute prices are no longer a barrier to IoT deployment. However, ‘Big Data’ is not so much about volumes of information, he argued, as it is about “let me find as much data as I need to find patterns in data,” a key consideration that should inform IoT implementations.
Terry observed that you can store anything these days (many varieties of data), but the trick is to know what you have, to secure it, and to be able to move it quickly. The key on this score is metadata, which provides control over the information and can solve issues in a digital library of assets, namely, the need for a catalogue of contents that would enable cross connect in 200 milliseconds or less, irrespective of the size of the data. A proper catalogue, as he described it, consists of an inventory of digital assets, (keywords, graphs), can deliver privacy-security-governance at the data layer (no more need for a VPN anymore), and can accommodate the continuous addition of structure/value to the data. What it delivers is a disruptive ability to return a query really quickly (Google can return a million results in .3 of a second), a prerequisite for the many IoT solutions that will be based on real time data analytics. For Terry, and many in the Wavefront audience based on their comments in the question period, the most critical issue is security and privacy – “how to do Big Data analytics with adult supervision” – and his answer was Privacy by Design, or “IoT with De-Identification.”
HP’s Daniel Gray also considered the data challenge in IoT, but ironically, from the perspective of “sparse data.” Field engineering VP for HP’s Big Data group, Gray has been able to tap into HP research resources in developing an approach to deriving insight from IoT data. Unlike Terry, who focused on ways to catalogue or filter information (reduce it to patterns), Gray argued the need for more data in IoT environments since the sophisticated types of analytics applied in this context – descriptive analytics, diagnostic analytics, predictive, prescriptive, and finally pre-emptive analytics – all depend on gathering increasing amounts of information, including historical data. To illustrate this point, he described a pilot study on the use of data in smart metering by utilities carried by HP Labs. In “Benchmarks for IoT,” researchers considered data requirements for demand management systems designed to optimize energy consumption, manage demand, reduce outages, improve operation, monitor resources and ultimately, maximize margins. The HP Labs Smart Metering Technology Proof Point analyzed 130 months’ worth of data, from 40 million smart meters and 22.8 trillion reading – a total data set of 726 TB – as well as a large amount of historical data. For Gray, several conclusions fall from this type of research: since load time for this massive volume of data is significant, supporting infrastructure and analytics engines must be up to the task (the Lab used HP Vertica, Big Data analysis, ANSI SQL engine and 8 x86 servers), and analytics capable of filling in the missing pieces in the case of sparse data sets, an apparent issue with IoT information.
Industry issues – IT and OT integration
“Has the Internet fundamentally changed everything as Bill Gates claims?” Richard Soley, CEO of the Object Management Group and executive director of the Industrial Internet Consortium, answered his question with a resounding ‘no’, noting the fact that disintermediation has been totally absent in areas like discrete manufacturing. In fact, Soley explained, there has been a distinct lack of “internet thinking” in manufacturing, and a related lack of interaction between IT and OT systems over the past 35 years –the same programming language is used in 2015, for example, to program a 984 programmable controller as was used in 1980 to program a Modicon 584. Similarly, there has been little change from 1950 – 2015 in energy grids, as in both eras, power is delivered from central point to millions of users using the same architecture. The only difference, according to Soley, is that now a very small amount of information is also delivered.
But this does not mean progress is not possible or even unlikely: for Soley, combination of the Industrial Revolution (concerned with physical means of production) with the Internet Revolution (disintermediation) equals the Industrial Internet, characterized by cyber-physical systems, M2M communications and new IoT models. One industry that is topical in this regard is the automotive sector, represented by the ‘connected car’, which has received a lot of attention for its ability to act as a mobile hot spot that links drivers to other services, and the vehicle itself to the manufacturer and maintenance businesses that will have the effect of improving efficiencies along the automotive supply chain. As Soley noted, currently, the car manufacturer becomes divorced from its product once the vehicle enters the distribution channel, but armed with lots more telemetry and other data and IoT communications, the OEM (and others) may develop new ways to improve customer service and develop new revenue streams. “Can manufacturing move to a subscription model?” Soley asked. The auto industry will be one of the first to encounter the huge social disruption associated with disintermediation, he argued, since with Internet models is likely very easy – as demonstrated by the phenomenal success of companies like Uber and Lyft – to improve on the average utilization rate for cars, which now sits at 6 percent. In this IoT context, how many people will want to own how many cars, Soley pondered.
Of course a key issue remains, which is the creation of standards that allow IT and OT (operational technologies) to talk to each other, and that can encompass the multiplicity of communications standards that might potentially be encountered in IoT deployments. As Soley put it, “the great thing about standards is, there are so many to choose from.” Ultimately, he explained, though there is competition among different groups promoting different networking protocols, etc., there will be no winners. Instead, Soley argued that there will be interoperability built into IoT solutions for specific industry contexts, several of which are currently underway in IIC Test beds managed by partners such as Bosch, Tech Mahindra, Cisco and National Instruments, to develop functionality for areas like security or command and control.
The third wave – sensors
If presentations on the Industrial Internet, data analytics and new business models described above offer some sense of IoT complexity, they by no means exhaust topics covered in the Summit’s track two. On the sensor front, for example, Guylain Roy-MacHabée, CEO of RX Networks, discussed the value in “location aware IoT,” where device location information can delivered via WiFi, cellular and GPS for integration into geo-aware apps, for interior spaces in particular, in addition to the communications challenges posed by the fact that devices may be fixed, mobile or nomadic, and efforts to create reliable reference points. Device security was another huge discussion topic that was addressed in Summit presentations as well as on the Innovation Showcase floor: Zighra, a startup that works out of the Invest Ottawa incubator, provided good illustration of the kind of innovation that is building in an evolving IoT ecosystem. The company’s adaptive behavioural authentication engine for smartphones works through gyroscopes and accelerators embedded into devices that can identify phone swipe patterns which in turn become unique identifiers – a type of biometric, but not fingerprint, approach to device security. As Zighra data scientist Colleen Smith explained, the challenges that remain – the potential to shut someone out of their device because that day they are doing a swipe that is slightly different or the possibility that an individual could mimic another’s swipe – are largely addressable through refinement of solution algorithms, though Zighra noted the lack of sensors in some of the cheaper, bare bones smartphone devices.
For more on IoT complexity, see the Wavefront Summit 2015 video compilation below, and stay tuned for a deep dive into the networking component of the Internet of Things.