In Part 2 of our Insight Vector series, Momenta’s Jim Fletcher provides perspective on market challenges, the different merits of horizontal versus vertical approaches and drivers for Industrial IoT adoption. IoT value needs to be obvious to end users, and this has been a lot more apparent in consumer markets, where there were not pre-existing technological alternatives. There’s plenty of value from horizontal platform to manage connectivity and data, but it’s domain expertise that delivers the most relevant business impact from Industrial IoT, and partnerships are key.
|
|
Jim Fletcher |
With extensive experience in executive management and technical leadership, Jim has held positions in areas of networking, IT systems management, and pervasive computing. He has a wealth of knowledge in IoT solution development including device connectivity, data collection, cognitive analytics, digital twin, and machine learning. He was most recently Distinguished Engineer and CTO for IBM’s Watson IoT Platform and is an IBM master inventor holding over 60 patents.
Q&A
What do you see as key challenges in the market?
One of the challenges of IoT is that the value needs to be readily apparent in the domain of the user. IoT cannot simply be 'change for the sake of change.' The iPhone was a dramatic change from any of its predecessors and began its dominance as a consumer-focused device. Few enterprises even considered the device early on. The device was so easy to learn to use and there was a complete ecosystem driving people across all socio-economic groups to consider the device even though most everyone would agree that is relatively expensive. The device gained acceptance and demand because of ease of use and perceived value.
Most of the IoT platforms in the cloud today are developed with IT mindsets for IT people. Their terminology, interfaces, configuration, and operations are IT-like, but their users each need IoT for some domain-specific use, whether it be consumer, or a wide range of industrial applications. Initially, all of the Platforms were focused on moving function and data to the cloud, but luckily there now is a strong awareness of the value that Edge computing adds to so many IoT use cases. In any domain, the solution needs to be targeted at its consumers using terminology they understand and driving value that is apparent to those that will utilize the system. There must be domain specificity and without that, you get a general platform without context.
IoT is about business transformation driven by the data you are collecting and the insights you can derive from it. Uber is a new business model for an old business. The values of Uber were enabled by connectivity, analytics, and changing the interaction model with the client to enable new perceived value.
Could you compare vertical versus horizontal approaches?
If you look at the anatomy of most IoT solutions there’s a lot of commonalities – you can connect, manage and analyze data. Once data comes in, you have in-flight analytics, based on what you learned from prior data, and then you create an efficient repository to build out models to be more aware so that the next generation has more intelligence. This is a never-ending cycle, because you can learn more and hone in on what my problems are. This is an area where edge has advantages and challenges.
However, if I isolate data to the edge, I need to be able to compare data at different edges. If you develop your models based on a single edge device, you won’t get the context to build effective models. This is where a horizontal approach is valuable.
Domain expertise is where context comes into play. Vertical differences in IoT are particularly pronounced in industrial – a Caterpillar tractor and a Honeywell BMS are very different. This is where you take horizontal platforms and get real insight.
Horizontal companies view partnerships as critical for domain expertise. We are starting to see OT people talk about things from the perspective of tradition protocols, looking to transform domain-specific edge protocols like the BACnet standard to JSON over Internet. If you think of the networking transition, SNA over IP is not how things will go; no one asks what network protocols are anymore. It’s not “IT for Things” - its IoT!
What areas do we see real progress in IoT?
Industrial continues operating with protocols and techniques that have worked for years, but for smart home and connected appliances there was nothing there, so totally new techniques have been developed. If you look at the interconnection that’s happening with Amazon Alexa – it can control smart bulbs, thermostats etc because through the web there are newly defined APIs to easily connect seemingly disparate systems, and systems like OAUTH to securely interconnect them. Where IoT hasn’t had as much success is in markets where there remain legacy alternatives. IoT has seen the most success where it is used to drive business transformation and provide new revenue sources.
Many Industrial IoT systems are inhibited by organizational boundaries and legacy procedures. A Water company didn’t want to connect SCADA systems to a network because of security concerns so instead– they dumped data over the wall to analyze two days later. One customer monitoring bridges had SCADA data that could be used for insights, but the OT team wouldn’t share, so the IT team had to build out a parallel sensor network to get data which was already available.
Can you share any books you’d recommend?
One of my favorite books is Future Crimes by Marc Goodman – it drives home where we are in relation to a digital world. If you think Google is giving us free products to use – their value is US, the people using the products. It’s an easy read and it opens your eyes.
To learn more about how IoT can unlock value within your industry, contact us at Momenta.