LEF perspectives: IoT paradigm shifts and the shift to the edge

The internet of things (IoT) could be the biggest thing to hit IT since the cloud, and it won’t leave IT looking quite the same. Business operational technology (OT) — such as machine sensors and telemetry data that was previously consumed by supervisory control and data acquisition (SCADA) systems — is increasingly being instrumented and enabled through IoT software, which very much sits within the IT realm.

IT is becoming business technology and with that comes more scope, more work, and some different thinking and approaches. Now IT is even more critical to the organization’s success.

The next paradigm shift involves IoT’s close relationship with machine intelligence (MI). One without the other makes little sense. Although MI brings its own sets of challenges — namely, in skill acquisition — it brings a significant advantage in insight that allows a business to act based on richer pools of data. The impact to business technology, not insignificant, will be what we’ve seen before: requirements for new skills, new tools and new partnerships.

At the same time, we see a race to the edge. Using centralized cloud-based models to collect, store, process and develop insight from vast quantities of IoT-generated data isn’t cost efficient, and in most cases is cost prohibitive. An edge-based architectural paradigm needs to be embraced, with a layer of MI, analysis, preprocessing, and data-wrangling done as close to the source as possible and prior to transmission to the cloud.

Once this cost efficiency is realized, you will get the added benefit of speeding data processing and be in a better position to deal with regulatory issues by storing data in the location it was generated in, rather than transferring it across borders. Data also has a value, which can rapidly decay over time. If your data is subject to fast levels of decay, then you need to analyze and respond to it quickly, making the edge the best place to do this.

But this isn’t a zero-sum game. The edge will augment the cloud, and a hybrid model is the best way to support it. However, the edge will also offer operational challenges for integrated monitoring and management, as well as for availability and reliability in the face of distributed partial failure realities.

That said, hybrid IT provides the ability to react in real time to data at the edge, while also aggregating data for analytics and integration with specific edge technology devices. Key data is then passed to the cloud for historical analytics, deep learning and training, long-term persistence of important data, and integration back into corporate systems.

Continue reading the position paper, Transform to a modern enterprise with hybrid IT.

Leading Edge Forum (LEF) is DXC Technology’s independent cross-industry think tank.