Wednesday, 15 April 2009

Green is Green

Energy costs are fast becoming the largest data centre operating cost, whilst up to 85% of computing capacity sits idle in distributed computing environments. Datacentres are physical structures that require tremendous amounts of power and cooling in order to operate and it is easy to see how improving this can translate into “greenbacks” for the organisation.


Baseline
From my perspective there are several key areas that organisations should tackle, but first and foremost, it is critical for themto implement a system to measure and track energy use. This will provide the baseline for driving a strategy to reduce energy consumption; reduce the impact on the environment; and improve operational efficiency.

There are some key enabling technologies that should be considered as part of an overall energy efficiency programme; and those that most impress me the most are virtualisation, cloud computing and purpose-built data centres.

Virtualisation
Virtualisation, both server and storage, is probably the fastest and easiest way to drive energy efficiency. The concept is a simple one: Maximise available resources (processing and storage), whilst “switching on” new resources when they are required. The separation of virtual machines/resources from physical machines/resources is implementable today and the business case proven.

Cloud Computing
Cloud computing is an interesting element in this overall story. Whilst we are still in the early stages of cloud maturity and adoption, it is definitely worth investigating. A cloud is essentially a pool of virtualised computer resources that we access as services.

The cloud platform is green by design as it dynamically provisions, configures, reconfigures, and de-provisions servers as needed. This separation (as with virtualisation) of what we want to process versus where it is processed provides us with the opportunity to maximise processing efficiency and drive green computing.

Data Centres
Now onto the actual data centre itself as the benefits of scaling a datacentre diminish if it is too large and requires too much energy.

There is an optimal efficiency that can be reached in a datacenter by matching the machines to the building (or vice versa). We can attain maximum efficiency by optimising all of the components that reside in the physical building, and monitoring power and heat with sensors that feed back into the systems management capability.

We can take this one step further by designing data centres with the type and load of processing in mind. This is the concept of a purpose-built data centre and not the current paradigm of adding servers and cooling to an existing building. There may be even be a variety of different types of datacenters, based on their respective purpose, within an overall network of processing capability.


So green is green; from an eco-friendly and economic perspective.

No comments: