How To Use The Green Grid Maturity Model

The Green Grid, the best-known group devoted to making data centres more efficient, announced a model earlier this year to help users measure how sustainable their data centre is.

Despite a briefing on the model,  eWEEK Europe was a bit sceptical. So we decided to look again, with the creator of the model, Harkeeret  “Harqs” Singh, head of energy and sustainable technology at Thomson Reuters.

Who should use it… and how?

The model is presented in a White Paper, which suggests users can get a “graphic equaliser” view of their data centre’s status, as shown below. But how do users produce these graphics, and what should they use them for?

“The Model takes you away from looking at sustainability as a project,” said Singh. “Energy efficiency is really a journey, and there are things you can do across the whole data centre. The model can show what your next step looks like.”

But it is not simple, because there are so many different ways in which a data centre can be adjusted and improved. Different organisations will set different goals, and focus on particular areas, depending on what their business is, what equipment they have to start with, and where they are in the world (because climate makes a big difference).

“Where you go depends on you and your business,” said Singh. The model will show senior management where they are today, and where they can go.

As he explained the “equaliser” diagram it became clearer. The section shown here shows only 20 of the columns – there are potentially around 47 contained in the model. These are grouped into areas, some of which apply to the building or facility  (power, cooling, management, lighting  etc) others apply to the IT systems (compute, storage, network etc).

In all those areas, the paper does not specify actual technology, because that is a business decision: “We shied away on purpose from being tech specific,” said Singh. In the facilities sections, users should reference their local building regulations, such as LEED in the US and BREEAM in Europe.

Level 2 is today’s best practice

The first thing he points out is that Level 2 is not average – level 2 is good. “Most typical data centres are between level 0 and level 2,” said Singh.  “Level 2 is current best practice, and Levels 3 to 5 are a guide for the future.”

These higher levels are really intended to help users to push designers for the future. “You use the higher levels in RFPs [requests for proposals sent to providers]. We wanted to get designers away from just re-using the last design they did.”

“Designers tend to be engineers,” he said. “If a design works, and no one has complained about it, they leave it as it is. We want to separate the designers who are innovative from those who are standing still.”

Level five is where data centres should be in five years’ time, he said, with levels 3 and 4 defined as intermediate steps towards that.
But what do the colours mean? White squares are undefined – there is no goal set in the paper, for level 5 in facility power efficiency (column 1.2) for instance. So users have the option to set their goals as high as they like below any white squares.

The image we have here is an example of a particular (theoretical) data centre, which has measured its achievements (in green) and set itself targets (in yellow).  The black squares are possibilities the organisation has decided are currently beyond it.

“Sites will set their own targets,” he said. “for instance, if a site is being closed, the organisation will not invest in it.”