As IoT Architectures Evolve, IoT Gateways Live on — but for How Long?
Nevertheless, some enterprises are using IoT devices to solve problems close to the edge of the network. As Cradlepoint’s Hosac put it, “If you look at this as a device evolution, gateways are turning into routers, and routers are turning into edge compute resources.”
He explained that systems which began with the focus of running simple scripts and binary files through their own hardware processors have evolved to include container-based networking capabilities, allowing them to interact with hosted clouds, “and basically push down machine learning algorithms to the edge” to enable local decision-making, Hosac added.
That evolution has much to do with how widely IoT architectures have grown in recent years, from isolated proofs-of-concept to include many more data-generating endpoints across LANs and WANs. But, it also has a lot to do with the emergence of IoT applications that can benefit from more edge computer power for processing, and that don’t require every bit (or byte) of data to be sent to a far-flung data center or cloud, according to Carsten Baumann, director of IIoT grid strategy at Schneider Electric.
At the heart of that application evolution is growing acceptance among enterprise customers to allow once-discrete OT and IT environments communicate, Baumann said. For example, industrial process automation and control applications traditionally have used PLC controllers. As more IoT sensors have come into these environments, IoT gateways have been used to collect and forward data from those sites, but remained discrete from the more traditional OT architectures. “Back then, there was a clear mandate that the IT and the OT worlds were not to interface with each other,” he said. “Now, because of new analytics and software capabilities, we are seeing in many different industries understanding the synergies and the efficiencies that can only be unlocked if you somehow bring the IT world in the OT world together.”
The benefits of doing this, as Partain, Baumann and Hosac acknowledge, include lower cost of bandwidth because less data is going back and forth between the customer premises and the cloud, as well as faster decision-making and the ability to act more quickly on data, perhaps automatically, because distance to the cloud or other organizational resources is no longer a factor. Also, multinational enterprises that previously had to be conscious of not running afoul of international regulations governing the transfer of data between locations in different countries may have an easier go by leveraging more edge processing and analytics capability.
As a result of this shift, IoT gateways are losing their top billing in IoT architecture to new kinds of platform — edge routers or converged edge systems, to borrow vendor marketing lingo — that combine networking, IoT management, processing power and security, and even software-defined networking.
Cradlepoint’s Hosac said a major aspect of the evolution beyond traditional IoT gateway has been to routers that don’t just include connectivity, but employ SD-WAN to allow dynamic use of connection paths. “You might have a police car that has a dual-modem router with one router pointed to Firstnet and the second one on Verizon, and then you can set application-specific policies for which network gets used based on certain performance characteristics or other factors,” Hosac said. The addition of such advanced capabilities to edge routers is part of the reason Hosac believes that the demand for traditional gateways eventually will dwindle.