Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Data centre directions 2014 roundtable: Reinvention and renewal

Jennifer O'Brien | Feb. 11, 2014
A group of select attendees gathered at an exclusive roundtable lunch to discuss data centre predictions for 2014 – with a key focus on the software-defined data centres promise and vision.

In the data centre of old, you had a lot of physical elements and, as part of our strategy, we came along from a historic perspective and said, 'well, they're going to abstract first and foremost that thing called the CPU.' And by moving that into a software construct we realised so many benefits, and many organisations have realised a ton of savings on the power side, and realised efficiencies they never thought were possible. We've been doing that now for 15 years, on average, and it's had a significant impact. To take that approach of what we did to the CPU and to extend it to other elements of the datacentre is what SDDC is all about.

An abstraction is that first key element to getting to that point. It's not just the abtraction of compute, it's also the network. The next shiny object on that space is network virtualisation — the traditional term of software defined networking where you're doing exactly the same thing. You're taking that traditional physical rigid inflexible construct and you're moving it into software. Because as soon as you move it into software you get agility, you get that scale, but you get that control of scale. We don't just stop at the network - we also move into storage. Having this whole premise of software defined storage and bringing the ability to get high availability and high-performance but from a consolidated and highly virtualised infrastructure is another key step in that journey to get to the SDDC.

JO: How is the SDDC solving the challenges associated with the traditional data centre silos?
AS : One of key problems in data centres today is siloed IT, where you have a complete silo for one application and you've got a complete silo for a different pipe. You run out of capacity in one; you can't just pick it up from another and give it to that to solve your service level issue. Being able to take those disparate resources across many different towers, put them into a large bucket and re-carve them out based on the service levels the business requires is the next key element to be able to realise the success of the software defined datacentre. Last but not least is to automate it to drive that efficiency. And you won't realise efficiencies around everything from right sizing to getting visibility into critical applications to being able to consolidate even further and to get that flow on effect around power saving and real estate savings, without automating it. Automation is an opex play: It's how you actually drive efficiency into the way your people manage their datacentres.

 

Previous Page  1  2  3  4  5  6  7  Next Page 

Sign up for CIO Asia eNewsletters.