Episode 46

Until recently we could perform Internet of Things computations in four general areas: We could compute in an external cloud, which means on one or more servers in a data center somewhere remote. We could compute “on prem”, which means on one or more servers in the enterprises’ local network. We could compute in the fog, which means on a gateway in the OT (Operational Technology) network or on a router or switch or some other network node in the IT (Information Technology) network. Or we could compute within in the IoT device or product, which means on an on-board embedded device.

We are now seeing a new class of computing surface emerge called the Mist. The Mist consists of the edge, that is, the very edge: the sensor and actuator controllers. Extending computing all the way to the edge can make a lot of sense depending on the network topology needed. For OT networks that are lossy and/or have low bandwidth, performing aggregation, fusion or filtering directly on the sensor could be handy. And since communications take 5x the power of computing, Mist computing may be a great fit for low power situation where extending battery life is of paramount concern. My favorite use case however is to use the Mist for parallel computing – solving big problems by breaking them up into smaller ones on smaller processors. Although more sophisticated to code, using a strength in numbers approach should be considered. While all of the five options are theoretically sound, deployment requires other variables such as overall cost to be factored in.

Here’s What We’ll Cover in this Video:

Watch this video to see Jurgo Preden discuss the state of the art in Mist computing.

Mentioned in this Episode and Other Useful Links