The rise in energy costs over the past few years is focusing the attention of companies and commentators not just on the amount of energy being used by IT, but more importantly, on the amount being wasted.
This situation, together with high profile campaigns by governments and other institutions addressing data centre energy efficiency, has driven a large number of data-centric companies to implement virtualisation strategies to reduce the energy consumption and increase the efficiency of the facilities under their control.
“The simple fact is that companies have not been able to realise the full financial benefits of virtualisation and in many cases have seen the energy efficiency of their facilities dramatically worsened,” says Rodney Callaghan, MD: Southern Africa at APC by Schneider Electric.
This situation can be avoided at an early stage and accordingly, APC suggests a simple five-stage approach to planning for virtualisation.
First of all, a company should understand the ways that energy is usefully and wastefully employed in its data centre to power, cool and protect the IT equipment. The EU Code of Conduct on Data Centres Energy Efficiency states that the data centre should be treated systematically in order to produce Power Usage Effectiveness (PUE) near to ideal.
Secondly, it’s useful for an organisation to have an idea of its eventual destination. To help those planning for change in their data centres, a number of simple and free to use TradeOff Tools have been published on the APC website.
“Thermal shutdown of IT equipment is an issue for those managing spaces using legacy cooling systems with perimeter CRAC and under floor air distribution, because such systems are neither predictable nor efficient in dynamic environments. The problem is magnified by the increased mission criticality of the physical equipment hosting the virtual machines.”
Subsequently, the next step for organisations should be considering cooling methodologies which are predictable and efficient and which can automatically respond to variations in power density so that cooling precisely matches demand.
Next, it is key to ensure scalability in terms of power and cooling. The reduction in IT load as a result of server consolidation offers the opportunity to take advantage of state-of-the-art modular, scalable architecture for power and cooling. With virtualisation projects, such solutions allow physical infrastructure to be scaled down to remove unneeded capacity without losing the option to scale back up again as the virtualised environment repopulates.
Finally, effective capacity management uses automated intelligence and modelling to monitor power, cooling and physical space capacities at room, row, rack and server level, suggests the best place for adding equipment, predicts the effect of proposed changes and recognises conditions or trends in good time for corrective action to be taken.