This page contains a Flash digital edition of a book.
FOCUS CLOUD UPDATE


Issue 11, Aug/Sept


THE CHANGING FACE OF THE DATA CENTER How much infl uence does the cloud have in the way power is channelled around the data center?


we have evidence that cloud is so different a compute load that a different power, cooling and organisational topology is required for it.


T


New ways of channelling power around a data center are, of course, necessary, but is it being driven purely by cloud requirements? Some believe the answer is not cut-and-dried. Much is made of the difference between a shared-use facility, such as a collocation site, and that of a data center dedicated to a single company and the services it provides.


One accepted difference is that neutral or shared colocation sites are designed to 1kW, 1.5kW or 2kW per square metre. Some will even stretch to 5kW per rack, but this is rare. Those designed for pure cloud environments are said to be capable of much higher densities; for example, Microsoft’s Dublin site was specced at 5-10kW per square metre capacity.


The power varies and the load is dynamic, moving to whichever section requires it, so the cooling provision needs to match the dynamic power changes to support it.


For example, by having homogeneous hardware, one can design efficiencies that are not possible in shared data centers. Thus, Google is said to be using DC supply to the rack and on-server mini DC batteries instead of risking even the 10% AC/DC conversion leakage and heat loss that is associated with traditional, three-phase AC uninterruptible power supply (UPS) systems.


DIFFERENT STROKES


A cloud service provider will offer different things. “If it is HPC, you’ll need higher power density than storage, which would be a medium-to-high power density, and then regular applications could require low- to-medium power density,” Chris Smith, marketing director at on365 said. “The provider (depending on which markets) will have to decide on the right fit.”


44 www.datacenterdynamics.com


o what extent is cloud computing changing the basic power-tree setup of the average data center? That question only makes sense if


Some argue that cloud refers to virtualization – which is surely at least part of the cloud story. The traditional data center is on the back foot here and has to be re-organized to do this more efficiently. The problem is density.


“Space is not the issue, but shifting power in and heat out is,” argues Aydin Kurt Elli, chief executive of business ISP Lumison. “With virtualization you need higher IT/power utilization per square metre, and legacy data centers are just not capable of running at high enough densities to support this going forward.”


Without doubt, there are some bleeding-edge centers with a mixture of high- and low-power density areas, with load changing in areas (possibly housing cloud nodes) so as to track a fluctuating level of (virtual) resources. “Data centers are being built with tiered power zones – with some parts being dedicated to low- density power consumption and other parts dedicated to higher density,” Jeremy Wallis, UK systems engineering director at storage firm NetApp said.


This means that more power can be dedicated to higher-demand applications, with other tiers being switched off entirely if not being used. This management system is a more efficient way of ensuring each application is catered for with as much or as little power as necessary.


NEW AND IMPROVED


Newer data centers are better equipped with hot- and cold-aisle containment, environmental monitoring and better methods to improve heat distribution, all set up to make cooling as efficient as possible.


But there are many reasons to want to build a new data center such as this – and that doesn’t mean it’s simply because you need it that way to run cloud applications. What they do have – think container, or in some cases not even that


ambitious for example, point of deployments (PODs) or halls within the raised floor space – is a build-out approach that promotes shared infrastructure for the power supply.


Colocation and cloud outfit Savvis told DCD Focus that it sets up cloud data centers around PODs that allow it to “distribute the power and optimize the calculations around power and space” but that, again, is more for general efficiency than just cloud rollout.


“While we do have a mixture of high- and low- power densities, and we constantly metre the power, our data centers are designed for high utilization, thus the impact of cloud is not that significant,” David Shacochis, Savvis vice president of research and development said.


And, of course, who says you can’t have such load balancing in a traditional (non-cloud) data center? Many have mixed environments: a high-density cube area using a hot-aisle containment approach, combined with a medium-density space with in-row cooling which supplements the existing down-flow cooling for the low-density requirements.


Only mixed-density data centers are best for offering cloud services. “It’s only new data center facilities that are set up to support high-density systems with high-power draw,” Stephen Owen, enterprise director for ControlCircle, a supplier of managed services points out.


But perhaps this is simply telling us that the switchgear, PDU and UPS makers – and their software development teams – are working hard in product development of power infrastructure for the cloud. As we move to metred over provisioned power, owners and operators of cloud service data centers and collocation or infrastructure-as-a-service providers’ chief concern is with chasing the load. 


Microsoft’s Dublin data center


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60  |  Page 61  |  Page 62  |  Page 63  |  Page 64