Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

IoT becomes an on-ramp for edge computing

A few years there was a lot of talk about IoT platforms. Analysts even created taxonomies of the IoT platform landscape that they saw developing.

While there are certainly many IoT products in the marketplace, there mostly aren’t any drop-in IoT platforms that solve a range of different business problems.

In fact, IoT is probably best when not thought of as a platform because IoT isn’t the platform– edge computing is. IoT never looked like much of a platform in the consumer space. Devices such as light bulbs, sensors and cameras talked directly to a solution on a remote cloud service, generally intermediated by some sort of hub. Edge computing was and often is vendor specific. In any case, it’s usually pretty simple architecturally.

The case for edge computing

IIoT is always going to be more complex. Most architectures comprise of at least an IoT gateway positioned near sensors and actuators. This model allows for filtering, aggregation and real-time responses to latency-sensitive events. In practice, most IIoT architectures are still more complex with multiple levels of gateways and other types of computing, storage and networking resources.

What’s happened over the past few years is that many different industries have embraced edge computing, especially for workloads that don’t look a lot like IIoT. This hybrid style of computing represents something of a reversal of the widespread view that everything would inevitably head to a public cloud.

Edge computing can be viewed as the newest enterprise IT footprint. In a sense, it’s the recognition that enterprise IT is heterogeneous and can’t be addressed with a single approach. Edge computing can also be viewed as the new platform for many types of workloads.

Where does this leave IoT?

IoT dovetails with edge computing in two primary ways. The first way is that IoT is a workload using edge computing as its platform.

However, not all edge computing applications are IoT, although many have at least some IoT elements. For example, computing at distributed retail branches is a form of edge computing. There may be an IoT angle here that relates to the use of scanners, Bluetooth or RFID.

But that’s probably not the primary focus of the edge computing platform in this scenario. That being said, technologies similar to the ones mentioned are increasingly pervasive, so more and more workloads are likely to have at least an element of IoT to them.

Other edge workloads are more squarely in IoT’s wheelhouse. For example, chemical plants, refineries and similar types of facilities are increasingly making use of drones and computer vision to quickly detect problems and execute predictive analytics. The volumes of data involved can be enormous. For example, Shell reported that their robotic subsea inspection videos exceed 7TB in size. It’s sometimes desirable to save everything against a possible future need. But with inspections, it may not be necessary for all the data that corresponds to nominal operations.

IoT can also be thought of as an on-ramp to edge. After all, edge architectures exist for a reason. And that reason is often to ingest information about the state of the physical world so that businesses can take action. In the case of the robotic inspection above, that action may be to dispatch a repair team. In the case of the earlier retail example, it may be to trigger the resupply of an item that’s hit its reorder point.

Data can also trigger a direct physical response. One of the reasons that businesses distribute computing and storage closer to the edge is to enable real-time responses to events. By handling things locally, businesses can avoid the latency and network reliability issues often associated with transmitting an alert to a central server and waiting for a response.

Taking a look at the trends, the data ingestion aspect of IoT will only become more important. It’s not that data is inherently valuable. Rather, machine learning and other data science tools are making progress in turning TBs and petabytes of data into higher revenues and lower costs for businesses.

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.