“I’ve seen situations in manufacturing where ‘too much’ data is flowing from the robot on the shop floor, through the local network, to the cloud and back again,” says Sathianathan. “That’s not a good thing because, as manufacturing CIOs know, decisions need to be made instantly to be effective.”
Even if you don’t reach the threshold of “too much data,” the value of AI/ML — and automation in general — is largely determined by speed. And that’s the number one thing IT leaders need to know bolivia mobile database running AI/ML workloads at the edge: speed matters, and its nemesis, latency, can be a killer.
Let's consider this and several other realities related to AI/ML at the edge.
To reiterate all of the above, the value of IoT data – or any data in edge environments – is very often tied to the speed at which it can be processed, analyzed, and acted upon.
In most automation contexts, performance is measured in small fractions of a second. “Transferring data from a smart device to the cloud to run an ML model, and then transmitting the results of that model back to the smart device takes too long in scenarios where milliseconds of additional latency matter,” says Chris McDermott, vice president of engineering at Wallaroo.ai.
Whether your “edge” is a vehicle, a utility, an assembly line, or one of the many other environments where speed matters, the costs — both financial and otherwise — of data transit and latency are likely too high to bear. “In these situations, the fastest place to run AI is at the edge,” McDermott argues.
2. Edge environments are the hottest place to use AI/ML right now
According to McDermott, the biggest growth in AI use today will be in a variety of settings that can be defined as “the edge” — whether that’s the car factory or the car itself once it’s on the road. The same goes for home appliances, power plants, and a long list of other objects that are now effectively used as IT environments.