Data centres: Taking control of TCO

Ask an enterprise about the TCO of its IT services and you would expect a swift answer. However, this will most likely be at best an estimate of the true figure. While this is acceptable for an off-the-cuff answer, it becomes more of an issue...

Share

Ask an enterprise about the TCO of its IT services and you would expect a swift answer. However, this will most likely be at best an estimate of the true figure. While this is acceptable for an off-the-cuff answer, it becomes more of an issue when attempting to calculate the exact costs and value of IT services and the infrastructure that they rest on.

More and more, organisations are finding that their initial TCO estimates for data centres and the services that they deliver become increasingly separated from the true costs as time goes by. With TCO a key component of IT investment decisions, this lack of accuracy can have serious ramifications for organisations that may be basing their IT strategy on flawed or incomplete calculations.

From Capex to Opex:
The problems with TCO aren’t new, but they are increasingly important due to fundamental changes in IT costs. Previously, the primary expense for a data centre would be the initial capital costs of construction and implementation of IT equipment; management, energy and other ongoing operational expenses could be assumed as constants and so easily factored into TCO calculations.

Yet as more organisations use virtualisation and on-demand IT services to make more flexible use of their IT infrastructure, management and investment strategies also have to change to accommodate this more flexible environment. At the same time, energy costs have become more volatile: while prices rise, it is not in a controlled, predictable manner. Both of these developments mean that the Opex needed to both power datacentres and maintain IT services has become a much more significant proportion of overall IT spending.

Flawed Calculations Produce Flawed Conclusions
The net result is that basing your TCO calculations on a static view of the data centre environment will no longer give you anything close to a reliable picture of your cost base over time. Given the increasing importance of energy costs, any TCO calculation should include the energy consumption of each individual component of an IT infrastructure throughout the year.

These calculations should also take into account the shifting nature of energy costs so that enterprises are not basing their TCO for a datacentre’s entire lifecycle on out-of-date figures. At the same time calculations must take into account how the infrastructure is used to provide individual IT services, in order to ensure that the enterprise will know the exact effects on data centre costs as demand for services changes.

To do this, enterprises should be able to drill down to the smallest detail of the IT infrastructure so that they know the impact of changes down to the penny. Perhaps most importantly, TCO calculations need to be an ongoing process: where predictions are made based on models of varying IT service and energy usage and then tracked against to ensure data centre costs stay on course.

Without this granular level of information, enterprises will find it increasingly difficult to cost and manage their IT services. An inability to measure the true energy use of their datacentres will make enterprises increasingly vulnerable to fluctuations in energy costs: making it easy for a rise in prices to turn into a budget black-hole that cannot be tracked.

At the same time, planning the capacity and delivery of IT services will become increasingly difficult without the ability to accurately predict how IT infrastructure is affected. This in turn will make long-term IT strategy planning a much less certain process that cannot guarantee the benefits it brings to the business.

Focusing on the Future
Addressing this issue requires two key actions from enterprises that will have wide-reaching consequences. First, businesses must see the data centre as a living, flexible resource rather than a single fixed cost centre. Second, they must ensure that they are aware of all costs and the factors that influence them, such as the expected lifecycle of equipment and seasonal temperatures in and around the datacentre, down to the most microscopic levels.

They can then combine the two to determine both the TCO of the data centre as a whole and of all IT services it houses. Armed with this information, enterprises can calculate the effects of energy price changes and predict the impact of any investment in IT services, confident that they are working with all relevant inputs.

Enterprises should not see changing the way in which they view data centre TCO as a huge undertaking. It is instead simply a logical step in the evolution of IT, from being based on fixed, capex-intensive assets to increasingly becoming Opex intensive services. In this environment, not knowing the exact costs and impact of each IT service is as unthinkable as Toyota not knowing the exact cost of producing a single Yaris.

Posted by By Zahl Limbuwala, CEO, Romonet

Enhanced by Zemanta

Find your next job with computerworld UK jobs