NABERS for data centres: Winners and losers

 

The data centre industry verdict on Australia's new energy rating system.

The New South Wales Government revealed its NABERS energy rating system for data centres to an inquisitive audience at the Australian Data Centre Strategy Summit this week, with a fairly clear picture of who will lose or gain from the rating now far more certain.

NABERS for data centres offers an official accreditor-led benchmark for the energy efficiency of IT workloads (the IT Equipment Rating), data centre plant (the Infrastructure Rating) and a fully-functioning data centre (the Whole Facility Rating).

The IT Equipment Rating – measured five times over one month – is based on energy consumption as a percentage of unformatted storage capacity and total processing capacity (number of server cores multiplied by clock speed for all servers).

The Infrastructure Rating is effectively a 12-month measure of PUE (power usage effectiveness) – dividing 12 months of energy use for the total building by 12 months of use by the actual IT equipment.

The Whole Facility rating is a 12-month combination of the two metrics to measure efficiency as a whole for CIOs.

Industry stakeholders interviewed at the Summit were generally supportive of the NABERS effort, but clearly some IT shops will benefit over others.

The winners

The winners, in an ideal world, will be every IT operator that pays less for their power bills. NABERS should – in principle – drive all data centre operators to look to ways to reduce the power consumption of their cooling and associated plant.

Whether those savings get passed back to the business/customer remains to be seen, they might just as well be absorbed by rising power prices.

In any case, an ecosystem of “Green IT” consultants and a bevy of new solutions from data centre power and cooling suppliers are there to take advantage.

The ratings will also aid those forward-thinking IT operators that built or moved into energy efficient facilities before the ratings scheme kicked in.

Some, like the NSW Government and National Australia Bank in recent months, will use it as leverage with their landlords.

“The people we lease buildings from, we have reduced rentals if those buildings don't hit performance marks,” noted National Australia Bank sustainability and data centre manager, Glenn Allan.

The losers

Not everybody will be a winner.

  • Those who fudged their PUE

Attendees at the conference noted that many PUE claims of existing data centres are wildly misleading. Some have taken spot measurements during winter time, when there is less of a requirement to use mechanical cooling, to publish a low figure.

“There has been no uniformity because there's been no way to verify,” said Bob Sharon, principal at Green Global Consulting. “For the first time we now have a rating system under NABERS that will standardise that measure and actually compare apples to apples.”

The 12-month read will soon put this practice to bed and reveal who has been fudging their numbers.

"As soon as the playing field is level and everyone's measured equally... it can't come quick enough for us,” said Greg Boorer, managing director of Canberra Data Centres.

  • Small IT

Boorer and many of his co-location and managed services peers might now expect a return on their investment in power-efficient facilities.

But for the majority of the tens of thousands of Australian companies with server rooms in their ‘back closets’, the news isn’t as good.

“We're actually seeing some landlords right now in Brisbane that won’t allow you to put a computer room into the building,” said Mike Andrea, director of Strategic Directions. “As soon as you suggest it, say goodbye to the lease, you probably won't get renewed or you'll be asked to leave at the end of the lease term.”

Matthew Clark, director of energy and water programs at the NSW Department of Environment, denied that government had any agenda to force Australian businesses into the co-lo model.

Boorer disagreed.

“Absolutely there is a push to get facilities out of office buildings and into purpose built data centres,” he said. “By 2015, all Commonwealth agencies have to be in 4.5 star office buildings, and it’s very difficult to achieve with an ageing server room supported by office air conditioning in the basement.”

  • Not all co-lo is created equal

That’s not to say that all co-lo providers will benefit. Upon reading the material the Government distributed at the Summit, some attendees noted that co-lo providers might find it difficult to change the behaviour of their customers around running their IT more efficiently.

“I have to accommodate any piece of equipment [customers] throw at me,” Boorer told the audience.

Much of that equipment may be inefficient.

“As this tool will indicate, 30-40 percent of the performance of the data centre is going to rely on the client's behaviour and deployment strategies within that technical space,” he said. “If the customers aren't interested, if you don't have that control written into your contracts , if you don't have technical working groups within your co-location facility to help your clients, then you're going to struggle to be world class in terms of efficiency.”

Thankfully for Boorer, his customers have been goaled around “hitting the Commonwealth's overall goals of minimum PUEs by 2015” and have educated themselves around “deployment methodologies, containment, air flow, cabling, and [anything] to maximise the efficiency of the facility.

“It’s my facility but it's also our facility in terms of meeting the PUE outcomes.”

  • New data centres

The NABERS system also favours existing facilities over new ones, as a rating can only be applied to a facility that has run (and been measured) for at least 12 months.

“You can't achieve NABERS without a historic track record,” Allan said. “So Day One you can't get a NABERS rating. You can't walk in one day and 24 hours later come out with a NABERS rating.”

Mindful of this, NABERS has developed a Reverse Calculator such that data centre designers can guess at the efficiency of new builds. The Government hopes it will encourage them to design to an expected star rating to drive better builds.

iTnews raised the potential for new builds to advertise an ‘expected’, ‘projected’ or ‘equivalent’ rating, which has been the case for advertising Tier ratings in the past (for availability).

“In the office building market we did see a bit of that early on, but I think market forces put pressure on people that are making ambit claims for things that they don't really have, and you'll soon find the people that are doing the right thing put the hard word on those that aren't,” responded Clark.

“In that sense, the public record's a wonderful thing,” Boorer agreed.

  • Cloud?

The other loser in the NABERS battle might be from unexpected corners. Enterprise customers are seeking cloud providers that have built out scale (and assume it is infinite). In order to provide that scale, many cloud providers drop in large quantities of server and storage whenever capacity becomes restrained. They are likely to be penalised for running equipment that is powered on but not in production.

Who do you think will win and lose from NABERS? Comment below.

Copyright © iTnews.com.au . All rights reserved.


NABERS for data centres: Winners and losers
 
 
 
Top Stories
Coalition's NBN cost-benefit study finds in favour of MTM
FTTP costs too much, would take too long.
 
Telcos finally briefed on data retention details
Update: AGD offers list of data to be stored.
 
Qld Health hires short-term CIO, CTO
Ray Brown leaves after five years at IT helm.
 
 
Sign up to receive iTnews email bulletins
   FOLLOW US...
Latest Comments
Polls
Which is the most prevalent cyber attack method your organisation faces?




   |   View results
Phishing and social engineering
  67%
 
Advanced persistent threats
  3%
 
Unpatched or unsupported software vulnerabilities
  12%
 
Denial of service attacks
  7%
 
Insider threats
  12%
TOTAL VOTES: 561

Vote