There is more life on the edge. People who live on the edge take risks, dream big, and claim to get the best view. “Edge effects” in ecology can lead to greater biodiversity. Leading edge technologies promise more of everything–more power, more functionality, more value. Now, as the conversation around cloud computing and the Industrial IoT expands, we are beginning to hear more about the edge.
Previously we looked at fog computing, where off-cloud computing power is installed on a cloud-connected device to reduce bandwidth and conserve cloud resources. But that’s just part of the picture, according to Harry Forbes, at the ARC Advisory Group. The edge of the IoT, he says, is not a single thing, but rather an area between the device and the transport to the cloud that can vary widely according to application requirements.
In a recent blog, “The Evolution of the IIoT Edge“, Forbes identifies at least five different aspects, or “classes of edge.” Here is a brief summary:
Intermittent Connection – A significant number of devices and applications don’t need a full-time connection to the IoT. A level sensor in a tank making a cellular connection to the cloud may need to send its data only a few times per day.
Data Historian – Sophisticated data users often log real-time data from multiple devices to a single, on-site historian. By connecting this historian to the IoT, the data can be sent intermittently to Big Data tools for processing and analysis. The edge application in this case combines consolidation and intermittent connection.
IIoT Gateway – This is a dedicated device that consolidates data from a number of locally-connected devices, and provides a connection to the IoT. Forbes points out that the value of this kind of device is that it can be provided and maintained by a 3rd party, and improved and updated as the service evolves.
Fog – An approach to cloud computing where some amount of intelligent processing is done at the device level, before the data gets sent to the cloud. This could include message throttling, conditional logic, and even simple control. We discussed this approach here.
On-premise cloud – For larger users like telecom services, localized cloud computing can be installed at the edge of the cloud to do what is essentially fog computing, but on a much larger scale.
As with the Industrial IoT itself, the vision and reality of what is happening at the edge is continually evolving. Forbes says, “Five years ago the idea of deploying a (on-premise) cloud at the base of every cell tower would be considered madness. Five years hence it may well be the norm.”