Following on from the great article write-up by my colleague Scott Truchot on 5G, this article intrigued me (if for no other reason than the spectre of the new buzzword/term). Fog (in this piece) refers to something between on-premise and cloud, essentially services running either extremely close to the cloud or network edge - there to support devices that are highly distributed and require real-time, low-latency connectivity. A way to think of it might be a reverse CDN for IOT devices as an example.
One of the key aspects of the move to 5G that really interests me is the potential massive increase in the number of directly-connected devices, as a result of the order-of-magnitude greater support 5G allows, and the much lower latency the new standard offers. It's this that is the big step change rather than any increased bandwidth - 50 billion+ smart devices connected even by 2020 as an example figure. This increase in connected hardware - directly speaking to some hosted service rather than via on-premise proxying or aggregator - drives the need for intermediatory receiving platforms; of which Fog is the term used here.
It's the security implications of this that I think will lag behind the explosion of device-count. Not located in a core datacenter, a vast range of manufacturers and standards, extremely high numbers of devices, and a greater sensitivity to denial-of-service and side-channel attacks all need addressing quickly to realise the benefits the new models can provide.
Fog computing brightens prospects for secure edge computing