We love data centres. There is something special about them – the pristine racks, the modern servers, colourful cables and all that raw processing power. The steady hum of thousands of servers, CRAC units and battery packs is intoxicating. They are great, but the problem is that they are expensive, power hungry and we need more of them. In 2007, data centres in Western Europe consumed a whopping 56 TWh of power, but as automation and online adoption continues to rise, this is estimated to climb up to 104 TWh by 2020. Dwindling oil supplies and tougher environmental legislation mean that generating that extra energy will become more and more expensive. In the current competitive climate, the vision of skyrocketing energy cost – or even energy crisis – is enough to cause sleepless nights for many European CEOs, particularly if they have seen their recent energy bills.
Due to the nature of the beast, data centres are prime targets for energy savings. They are power hungry yet well metered and therefore, great places to start flexing those cost-saving muscles. Most data centre operators have existing monitoring systems in place. At a minimum, these systems can produce totals for utility and IT load, so baselines are there. Savings are easy to estimate, and therefore, the great machine is often asked to do more with less. This is by no means the only challenge data centre operators face today.
These are no idle thoughts. As concerns about our energy reserves grow, everyone should make it a norm to be as energy-efficient as possible. The European Commission Institute for Energy and Transport acknowledged this fact in 2008 by creating a voluntary EU Datacentre Code of Conduct, which, like all similar schemes, will involve submitting annual reports. The submission process will involve energy baselines, initiatives and action plans. Getting baseline data out of a traditional data centre should be easy, but identifying initiatives and formulating action plans could be a headache if the monitoring systems are not capable of producing detailed live reports on how the data centre actually runs. The action plans will most likely include things like virtualization and optimal resource utilization, but without proper monitoring systems it is very difficult to make the right move.
