Latest Posts

Latest Comments


Seeing through the fog


Posted by |

Do you remember when the cloud seemed like just a marketing term for centralised computing? Well, over time the term cloud was more clearly defined, and as the acceptance of cloud rose so did the adoption rates. Now the cloud is not just an accepted method for delivering IT services, in many cases it's becoming the default. So, the industry is at it again and this time it's 'fog computing'. Before we dismiss this new term as simply good marketing, let's take a look at what it actually means.

Fog computing is being used to describe an environment where data, processing and applications are undertaken by devices around the network edge rather than entirely in the cloud. As a concept, fog computing is being driven by two things: the increasing number of internet connected devices driven in part by the rapid rise in machine to machine communication (IOT); and the challenge of engineering centralized data centre networks that can cope with this rise in traffic.  

It's undeniable that the Internet of Things (IoT) is on its way to becoming a reality - from domestic smart energy meters, to your Sky box, the fridge, and even car, we're seeing new internet-ready products launched daily. And of course as we continue to adopt these, so we are driving the need for information to be processed and returned to the device for it to operate successfully.

Most of these smart devices however aren't that smart. Many are simply there to receive communications and directions and have little local storage. Complex Data processing would be beyond their memory and processor capabilities - or if they can compute data, then the price would certainly need to reflect the computer inside. However, as any large organisation who processes its data centrally but needs to send it to a device already knows, bandwidth can be a challenge. In a fog computing environment then you need to look to larger devices at the network edge, to fulfill the data processing and storage requirements.

However, these devices weren't developed to process copious amounts of data nor were they developed to keep it secure. While it appears logical to engineer compute, store and make decisions at the network edge, it's still important to consider physical security and that's harder to achieve at the edge.

Decentralising your environment also returns to a scenario where you need to spread your capital investment and resources more thinly. And if we re-divert data processing to the network edge there is a risk that the economic benefit of centralisation is negated.

So the question is, where on the network should we undertake data processing? Locally on the device, on the edge, or centrally in the cloud? The answer is, of course, all. If there's one thing that we know about customers, it's that one size doesn't fit all. Each need, environment, business driver and objective is unique, and managing your IT environment should be too.



Post a comment

Comment submitted! Comments needs approval before being displayed.