Data Centre Specialist Group

IT costs are rising out of control, driving us toward a financial crisis, at the same time the carbon footprint and energy efficiency of Data Centres is becoming a significant financial, regulatory and social responsibility issue.

The Data Centre, whilst being the single most valuable and expensive asset is frequently the last to be considered in IT strategy. The presentation at our event on the 19th September will demonstrate why M&E should have at least as much input to IT strategy as the IT and Architecture teams. See our event calendar for the 19th September members meeting

Rising demand for IT services is driving demand for data centre capacity, in response to this many companies have replaced old servers with physically smaller servers or blades, only to discover that their capacity limits are power and cooling, not physical space. These internal issues are compounded by the lack of availability of additional power in many locations.

The falling prices of server hardware have created an illusion of falling operational costs which is belied by the rising power demand and power density. The cost of a server is now matched by its lifetime power cost, the cost of the Data Centre infrastructure required to power and cool the server substantially exceeds both put together. The only way for a business to effectively plan an IT strategy and choose appropriate IT hardware is to understand the overall cost of housing, powering and cooling that equipment.

There is no shortage of solutions being offered to IT and Data Centre operators, from virtualisation to rack-side cooling but little assistance in understanding what real environmental and cost benefits these might provide you in your Data Centre, delivering your IT loads.

The Data Centre Specialist Group has developed a model to analyse and forecast the integrated IT and M&E energy and financial costs of purchasing, maintaining, powering and cooling different types of server. This model demonstrates that choosing servers based on price or price/performance is the wrong strategy. The performance per Watt and rated power of the server are the key metrics for lifecycle cost and should, therefore be used to select equipment. The model is also able to effectively forecast the carbon and cost benefits of migrating to a virtualised or grid environment, allowing effective analysis of when and if you will see the ROI.

Using the model we have been able to test the energy and financial impacts of changes in both IT and M&E strategy and processes. There are a number of immediately effective changes for which we can demonstrate the cost benefits as well as forecasting the medium term impacts of equipment, software, process, taxation and regulatory changes.

One significant output of the model is that the new breed of energy efficient servers will create a new set of M&E issues in the Data Centre due to high load variability. These issues will require a higher level of communication between IT and M&E staff both provisioning and operating the facility.

One of the most significant challenges facing Data Centre operators is communicating the impacts of IT and M&E strategy at board level. The model allows the IT and M&E issues to be communicated effectively in the business terms of capital and lifecycle costs, to allow business decisions makers to make effective decisions based on a real understanding of the choices presented.

The DCSG will be releasing the first public review version of our white paper, “Server Energy Efficiency and Power to Load Linearity, and the impact on data centre energy efficiency” to our members at this meeting. This paper explores the energy and cost issues facing a Data Centre operator today by describing current and future servers and comparing their lifecycle energy and financial costs under a range of Data Centre scenarios.