Think of a fog computing node as a bodily server that resides concerning the edge gadgets (thermostats, robots, in-motor vehicle computer systems) and the back-conclude devices, generally hosted on public clouds. The fog nodes react to an architectural difficulty: also substantially latency to go requests all the way back to the public cloud-primarily based services and not sufficient horsepower to course of action the data in the edge device itself.
This 3-tier program adds a further compute system that is in a position to do some—if not most—of the back-conclude processing. This addresses the problem that more affordable edge gadgets with reduced electrical power really do not have the processing and storage electrical power to deal with the incoming data natively. Now data can be despatched to the fog node for processing, without having the latency effect of likely all the way back to the remote cloud services.
Though fog nodes are a very simple resolution to a very simple challenge, you should really recognize when and when not to use them:
You should really use fog nodes when data that is sophisticated or in substantial quantities want to be processed regionally and would overwhelm the edge-primarily based device that is consuming the data as properly. In other text, you want a thing that responds in nearly genuine time, this kind of as a factory robot shutting down when the servers overheat. You want that to be instantaneous.
You should really use fog nodes when a human is waiting for the reaction to return back to the edge device. Like the want to react to events that just can’t wait around, people want to be deemed in the vicinity of-genuine-time architectural elements. People today are also high-priced to be waiting close to for responses from the remote cloud devices.
You should really use fog nodes when you are multiplexing data from quite a few different sorts of edge gadgets or sending quite a few different sorts of data. The fog node manages the processing of data from quite a few edge gadgets at as soon as, dealing with things this kind of as semantic transformation of the data prior to sending it to the back-conclude cloud servers. Or the fog node can course of action and react straight to the edge-primarily based device.
In essence, you should really not use fog nodes unless of course the architecture and demands in good shape the higher than requirements. In my reserve, they should really be utilised sparingly, especially thinking of that they include cost and operational complexity—and a further layer that can are unsuccessful.
Your most effective bet for now is to exclude them, and then figure out if you have conditions where by they can be of use.
Copyright © 2020 IDG Communications, Inc.