Marginal gains – what the Azure UK data centres mean for UK PLC
By: Siobhan Wood on4 minutes to read
Last month Microsoft announced a new data centre region in the UK. Whilst much of the focus was on what this means for central government departments, our developer looks at what the proximity of these new “Hyperscale” facilities means for companies who rely on speed for their competitive advantage:
Cloud computing. As nice an idea as it is to have our data gently floating by in the sky – as most people know (except maybe my Gran) this isn’t actually the case. The reality is that all of those servers live in huge, faceless buildings – generally, far away from civilisation in ‘low-risk’ areas. Just yesterday, Mark Zuckerberg shared a set of photos from Facebook’s own Arctic Data Centre and Microsoft have announced plans to build their new data centres under-water – a.k.a Project Natick
As an evangelist of Microsoft Azure, not just within my workplace but also personally, it always disheartened me when I went to create a new Virtual Machine for a client or a new SQL Server for an app we’re working on and my only options were to run these resources in either Dublin (North Europe) or Amsterdam (West Europe). I appreciate these are both still relatively close locations, but it always seemed unusual that there were no offerings within the UK.
Excitingly, all that has now changed with the official launch last month of two new Azure regions, now available publically. ‘UK West’ and ‘UK South’ were opened for business this month making Microsoft the very first hyperscale cloud provider to offer services locally within the UK.
You may be thinking that the difference in distance to either West or North Europe from the UK would be marginal – what benefit does this actually bring? Well in an age where even the smallest delay in data travel can cost companies millions every year, every millisecond certainly counts, so this is great news for service providers in the UK. It has been recently documented that Amazon estimated every 100ms of latency their site experiences costs them 1% in sales, so this should give you an idea of just how critical data transfer speed is.
The table above was pulled from the live view of global ping speeds that is available from https://wondernetwork.com/pings/. Clearly putting a data centre in Auckland to service users in Barcelona would be a bad idea.
If you want to understand more about the affect of latency across the internet and just how important a factor it can play, I’ve found these resources invaluable in gaining a deeper understanding:
- Latency: The New Web Performance Bottleneck – this is a brilliant explanation on the issues that latency poses and why there are no simple answers to improving it other than “shorter cables”
- It’s the latency, Stupid – something of a seminal essay, first published by Stuart Cheshire in 1996 and still relevant today
- Layers of Latency, Cloud Complexity and Performance – A 2012 article by Wired that explored the impact of latency on the then still nascent cloud industry.
The increased speed in data retrieval is not the only advantage that UK based data centres bring. Just the very fact that the physical location of the servers resides within the UK means their services are open to huge organisations such as the NHS or Ministry of Defence who have strict regulations on where their data lives. The MoD actually became the first client to launch services in the new region.
With the Olympics and Paralympics just behind us and with yet more success on the cycling track the issue of latency is definitely something that David Brailsford would put into the “marginal gains” category:
Currently, the resources available in the UK regions are limited but it’s safe to assume Microsoft will be rolling out more and more of its core services in due time and that offers new competitive advantages to lots of UK businesses.