Virtualizing waste – An expensive mistake?

06-03-2012 - John Hatcher

An increase in virtualised servers may be a success story, but if the growth is unexpected it is likely more attributable to virtual sprawl, one of the quickest ways to negate the benefits of a virtualisation project. When the anticipated growth rate of a server estate is exceeded a business needs to understand why.

Looking at a virtualised environment from an efficiency perspective reveals where anticipated savings are being lost across facilities and IT.

Hidden costs of virtualisation
Hosting multiple virtualised servers requires more CPUs and significantly more memory, both of which contribute to power requirements and heat output. If growth is not managed, energy demand will increase as organisations purchase more high specification host servers to support the growing number of virtual machines. This is a hidden cost for facilities departments.
In addition, initial savings gained from virtualisation will be lost if virtual machines continue running when no longer required. Research and advisory firm Gartner refers to this as a 'Virtualisation Catch 22' .

Make virtualisation more efficient
New efficiency tools can help identify what is useful and when, making it easy to establish a process for reclaiming unused resource and avoiding unnecessary spend on new resource. The Green Grid’s Data Centre Compute Efficiency Metric (DCcE) is a good example of how useful work can be calculated. The Green Grid considers a server has a primary purpose where all resource should ideally be used. Any resources allocated to secondary operations detract from the ideal and are wasteful. Often virtualisation projects do not include these metrics, limiting the benefits that can be gained in efficiency.

In addition to metrics, a company can also do the following to help realise the benefits of virtualisation:

1. Optimise the environment – Look at assessing efficiency as well as uptime and performance. Efficiency audits identify hidden leaks in terms of cost, wasted energy, unused software, inefficient hardware and legacy systems that should be decommissioned, rather than virtualised.
2. Set a benchmark for efficiency – Once the percentage of servers doing useful work has been identified, a benchmark for efficiency can be established making it easier to monitor for sprawl in the future.
3. Improve IT waste discovery – Be prepared to find waste. When communicating audit benefits back to the business, champion savings made or that stand to be made, in quantifiable cash, energy or tonnes of reduced CO2 emissions.

Visualise in order to virtualise efficiently
Presenting data visually offers executives transparency over what is useful, unused or being wasted. Continuous monitoring proves that a virtualised environment is efficient. It is a better solution than checks initiated by a capacity problem or reactive mandate.

Visualising that a server uses power for useful work all the time (left) and wastes most of the power doing almost nothing useful (right)

Publishing on-going measurements internally via a dashboard is also a powerful way of combating sprawl by modifying behaviour.

Chargeback schemes can help organisations gain transparency of costs. These schemes facilitate business and IT accountability by moving cash flow controls away from IT and back towards business where costs were initially incurred. This mechanism prevents and controls virtual sprawl, helping business units use only what they really need.

Virtualisation has clear benefits, including a reduction in energy consumption and floor space. To get the most out of virtualisation, companies should be prepared to invest in tools monitoring efficiency.

Creating visual representations of server workloads helps companies to efficiently identify virtual sprawl within their environment. Such visualisations ensure greater control of issues; facilitating the identification and elimination of waste, leaving companies to experience the full potential of virtualisation.