Edge means different things to different people

What is edge computing?

There’s a lot of discussion within the ICT world of edge technologies, and processing at the edge. But it’s not always clear what the 'Edge’ actually is. We’ve compiled five different definitions that are used at different times by different market participants. They are:

• For data centre providers, ‘Edge’ is about moving co-location space closer to users including establishing presence in secondary and potentially tertiary cities.
• Mobile operators have a concept of Mobile Edge Computing (MEC), which involves the co-location of cloud computing power at the (cellular) network edge (EnodeB, or RNC).
• The term Mobile Edge Computing (MEC) quickly evolved into Multi-access Edge Computing (also abbreviated to MEC), providing for the deployment of computing power at the edge of any network (including Wi-Fi access points, and so on).
• Currently, the most widely accepted definition of Edge relates to the piece of IT (Information Technology) hardware that is nearest to an OT (Operational Technology) asset, and acts as a bridge between the OT and IT worlds.
• To a software architect, the Edge is a software container that can potentially reside on an OT asset, like for example a video camera with onboard processing capability.

What do we mean when we say ‘Edge’? Well, all of them. The key concept of edge (that all of the definitions above have in common) is around moving processing power closer to (or onboard) end devices and further away from any centralised processing infrastructure. But what really matters is how all of the different ‘edgy’ locations are combined to support an application.

Delivering edge-enabled solutions

The key consideration should always be to use the edge location that is best suited to a DX solution, potentially delegating elements of application processing to ‘nearby’ or ‘upstream’ devices that have spare capacity. This is more often than not a trade-off between cheaper processing power (up the chain) and lower round-trip delays and the potential to maintain service during network connectivity outages (towards any connected end device). Another consideration is often the optimisation of wide area bandwidth use, with edge analytics (redaction at the edge) being particularly important in the case of CCTV monitoring, for example.