The move to edge computing is rapidly gaining momentum. Ensuring, for example, a totally immersive experience when playing computer games, navigating the metaverse, or performing remote surgery.
And allowing manufacturers and other industrial users to achieve greater efficiencies – by leveraging vast amounts of data in real-time from their super-connected plant and production facilities.
However, all the above depends on ‘blink of an eye’ response times which in turn demands low latency networks for enabling the transmission of all those data and images as fast as possible between devices and servers for processing. This puts a huge emphasis on the availability of internet exchanges, the data centres concerned being physically close enough to the network edge, the quality of their fibre connections, and sufficient power and cooling.
Fortunately for all those enterprises, manufacturers, cloud and content providers looking to get data and services closer to users, the UK’s regional fibre infrastructure, along with 5G, is advancing at pace. The major fibre providers, carriers, and metro network specialists are rapidly connecting to a growing number of strategically located edge data centres around the UK, allowing much greater accessibility to high-performance networks and greater bandwidth availability.
Equally pivotal are the internet exchanges for routing the torrent of data traffic now at stake. The traditional centralised model of internet exchanges in place since the mid-90s has served us well so far. But even with port sizes increasing from 10 Mb to 400 Gb, it is becoming unviable for all internet traffic to continue being managed via the current handful of exchanges.
In the new IoT edge computing world, there is simply too much data for these to handle and without taking a fresh approach ,this will impact increasingly on application and service response times as well as transit data costs. In the same way that the early ISPs realised that keeping traffic local drove down costs and improved performance, we now urgently need more regional internet exchanges to complement the existing infrastructure model.
Together with edge data centres, these can play a key role by keeping traffic local thereby helping ensure the lowest possible response times while at the same time preventing ISPs’ data transit costs from rocketing. The availability of these resources at a local level will also support digital businesses, regional tech hubs and bring significant economic growth to the regions. These will in turn bring new opportunities to the benefit of local communities.
Power and cooling
Together with more regional internet exchanges, abundant fibre networks and well positioned data centres, there is a further prerequisite for achieving the full potential of edge computing: data centre power and cooling availability. A continuing shift to decentralised edge computing is already seeing more hybrid cloud and HPC deployments, both of which have high density power to rack requirements.
An immersive technology application, for example, such as a 3D virtual reality modelling simulation will demand considerable power to rack to support it – we are already seeing as much as 25 KW in single racks. There are also the bespoke cooling requirements that many of the emerging IoT and AI-based edge applications require.
Therefore, to ensure optimal performance, CIOs, solutions architects and system integrators currently developing their next generation of edge applications and services will need to carefully evaluate a data centre’s forwards power as well as its proximity to fibre networks and local internet exchanges.
We believe regional edge data centres offering the necessary power and critical infrastructure are ideally placed to support the existing overloaded exchange infrastructure. This is why we have recently announced our intention to equip all Proximity’s current and future regional edge data centres with internet exchange facilities.