Recently we have seen the notion of “fog computing” emerge to describe the dispersal of cloud technology out to the edges of the network. This term has particularly been embraced by Cisco to describe a new management paradigm for the Internet of Things (IoT).
In this IoT model all the local devices and sensors deployed at the edge attach to local gateways, which connect through the service-provider’s access and edge networks to the cloud. These local gateways are becoming increasingly smart and sophisticated – driven in part by the security imperative (to avoid the global IoT descending into chaos:) .
With the Fog model, the cloud retains its central “think-tank” role (analyzing data and making all the big decisions). When there are no resource constraints and there’s a flood of data with linkages among multiple data sources, it obviously makes sense to centralize and perform everything in the cloud.
However the cloud can also delegate some tasks out to the smart gateways and access systems as it often makes more sense to localize analysis and decision making at the edge. For example when there are constraints on time the Fog model enables the IoT to deliver a quick response at the edge, without being burdened by network latency. Also while it’s not the mission of the smart edge device to undertake in-depth analytics, the device can actively filter local data and selectively relay data to the cloud (e.g. don’t transmit video of empty rooms) and the traffic savings for widely distributed sensor networks can be considerable.
So we see Fog computing will bring a whole new breed of applications and services … all built on cloud technologies but with selected applications distributed out to appliances at the edge of the network. May be companies like Opengear should start referring to their ACM5000 and ACM5500 smart OOB remote site managers and gateways as fog appliances.