Content Copyright © 2016 Bloor. All Rights Reserved.
Also posted on: IT Infrastructure
It would be tempting to think that the emergence of an Edge computing paradigm would be a God-send for those European data centres located away from the key interconnectivity hubs like London, Amsterdam, Paris and Frankfurt. While the Internet of Things (IoT) may indeed provide opportunities for these operators, the current drivers of Edge computing point to a very specific proposition that is unlike a general co-location offer.
The explosive growth of content streaming is the driving force behind the Edge computing story. Netflix, Amazon Prime and YouTube are huge users of bandwidth. As they compete for eye-balls with traditional broadcast media they naturally want you to be able to feast on full HD with no buffering. Delivering that sort of performance from the big Equinix and Interxion data centres is no problem if you live near them. But what happens if you are in Glasgow, Marseille or Munich? If you can only deliver content at standard definition, or there is too much jittering and buffering you very quickly see people move away from the page with an obvious detriment to revenue.
User experience isn’t the only issue. Transporting this content over hundreds or thousands of miles and relying on peering agreements with network providers can be very costly. What is needed therefore is a local facility where both the content distribution networks and ISPs can come together and interconnect. A recent study by analyst firm ACG in the U.S., paid for by EdgeConneX, showed that in a conurbation of 1 million people local caching could save about $110 million in backbone transport costs over 5 years…a saving of 50%.
So far so good for local data centres surely? Yes, as long as you are in a metropolitan area of 1 million people or more and you have been able to attract both the ISPs and the content providers. This isn’t as simple as it sounds. It is a real chicken and egg scenario that requires a highly focused sales effort with what are major international organisations; not easy for small local data centre operators.
The content providers also require high density facilities and assurance that new capacity can be brought on stream very quickly, where they need it, not necessarily where the data centre operator has existing or planned capacity. To give you an idea how quickly things have moved in the U.S. EdgeConneX has gone from zero to building data centres in 24 Tier2 cities in 2 years. This is a challenge that shouldn’t be under-estimated. How that can be achieved, and whether EdgeConneX will be able to match that sort of roll out pace in Europe is something we’ll come back to in a future post.
Perhaps I am a little cynical, but I fear that Edge computing as a term will come to mean something far broader than the very specific requirement and supporting proposition I have alluded to above. In the long term, a little like Cloud, it probably won’t matter and may even drive wider uptake of data centre services. However, in the short term, smaller regional data centre operators should be aware of the very specific requirements of the current Edge market and the not insubstantial challenges of meeting them. Location will be a differentiator. But it is not where you have a data centre but where the ISPs and Content providers want them that will be crucial.
This post first appeared on the old Cassini Reviews website.