Changing under the skin

Written By:
Published:
Content Copyright © 2015 Bloor. All Rights Reserved.
Also posted on: IT Infrastructure
Leadership search and business recruitment concept as a group of pawn chess pieces and one individual standing out with a king crown cast shadow as a metaphor for the chosen one.

At IPExpo last month there was a panel debate on the Future of The Data Centre which concluded that the data centre in 2020 will look pretty much like the data centre of 2015. At DCD Converged last week an impassioned question from the floor asked why so little change had been seen in data centres over the past 20 years and was there a danger of them dying?

The simple answer is no, they won’t die. In the short to medium term we’ll need plenty more of them and I don’t need to revisit the trends and facts that point to that conclusion here; they are well signposted. But do the questions posed at IPExpo and DCD Converged accurately reflect that significant changes aren’t occurring?

Appearances can be deceptive. The existing racks in data centres may be constructs of the early 20th, rather than early 21st Century, but the layout and management of racks has changed to deliver more effective energy management. The thinking and technology around cooling has advanced significantly, cabling and power management has become more flexible. Operations have become more automated; waste heat is being used more imaginatively. The whole idea of the software designed data centre and the emergence of Open Compute and Open Stack is changing the server, storage and networking equipment that goes into data centres and the way in which they are managed. Imagine transporting a data centre operations manager from 1995 to the present day. He would recognise he is in a data centre but he would probably be amazed at what was going on it and how it was being run.

But this hardly adds up to the disruptive changes we have seen in other industries. The long term capital investment required to build data centres and the huge overhang of legacy IT infrastructure obviously militates against disruptive new entrants. But this is not to say that it can’t or won’t happen. Many people envisage a time when all computing is carried out in the public cloud. If this comes about and the three big players, AWS, Microsoft and Google turn out to be the winners this could be very bad news for a whole range of data centre operators, but it doesn’t necessarily signal the demise of data centres themselves.

The growth of IoT, Big Data and Content Streaming are all pointing towards the need for a move to the Edge. In fact, if you listen to those with visions of the future everyone on the planet could become a data centre given that we all own extensive computing capability and that we are each likely to have multiple IP addresses. A little nearer in perhaps, the growth of modularised and micro data centres and the improvement in local network infrastructure leads some to believe that networks of small local data centres may be much more efficient than the huge buildings we see going up today. There are many challenges these visions will have to overcome. Not the least will be the provision of ubiquitous, flexible, resilient and secure bandwidth. As Peter Judge, the Global Editor of DataCenterDynamics, said at DCD Converged last week, “Compute and Storage know that Networking is the problem child.”

None of this points to a massive short term disruption of physical data centres. Change will come in incremental steps in the foreseeable future. Operators do need to be aware of the changes in the market and those that adapt their business models and value propositions quickly will be the winners. However, individual components may well be disrupted. Consider for a moment what IoT might do to the DCIM market; a series of £20 sensors communicating with each other and feeding into control and monitoring systems put together with open source software, against DCIM suites costing tens, or even hundreds of thousands of pounds to buy and implement. That could be a very short battle. Then again, will Vapor.io disrupt the traditional rack market? We shall see.

Just as someone awaking for the first time since the 1930s and walking onto Westminster Bridge would notice a lot of familiar sites, today’s data centres may look very similar to those of 20 years ago, but under the skin a lot has changed.

This post first appeared on the old Cassini Reviews website.