Air.  It's a wonderful thing. Little did 18th century researchers like Joseph Priestly (1733–1804) and Antoine-Laurent de Lavoisier (1743–1794), who investigated the properties of air, know it would become a cooling medium for information technology (IT) in the 20th and 21st centuries.

 

The mechanical technology that is employed to cool IT is well established.  And the data centre industry has learnt that the principles of cooling IT requires a little bit more thought than just blowing air in the void created by a raised floor. In the bad old days of the early computer rooms, the under floor was nothing more than the dumping ground for routing data cabling, power cabling and piping. The poor old air molecules had no chance and the IT suffered.

 

Today, there is a plethora of solutions to cool IT in a modern data centre. A primary design requirement, that data centre users have always stipulated, is that the system must meet tight guidelines for controlling temperature, humidity and air filtration.

 

To meet this requirement, the cooling systems provisioned in data centres have been 'closed'. By this we mean that the air in the data centre is not mixed with the outside air. This ensures that the aforementioned parameters can be controlled. There is tighter feedback within the system to smooth out the fluctuations and thus episodic atmospheric events occurring outside the data centre do not impact the data centre environment and its IT.

 

Manufacturers of IT equipment and leading industry bodies such as ASHRAE have codified the requirements to cool IT equipment and this has lead to significantly better science in the solutions that vendors are now marketing.

 

One area that absolutely confounds me though, is the debate over air-side economisers to cool the modern data centre. It's not the fact that they can be used—this is being demonstrated in a very practical manner in latitudes from 30˚ and higher, and the debate is now focusing on even lower latitudes where the use of the air-side economiser may be less, but over a longer period it provides benefits.  What confounds me is the fact that data centre users are being wooed by mainstream mechanical engineers who say they can construct an air-side economiser from scratch (using common equipment and controls) and meet the quality control requirements of a closed cooling system.

 

The data centre is not an R&D tool for mechanical engineers. It's an operational environment providing critical services to the owner's business and its customers. Yet, data centre users are willing to give it a go, based on a mechanical engineer’s 'trust me, she'll be right, mate' approach.

 

Has there been a major positional shift in the risk associated with this data centre element? Have I missed something? What happened to maintaining the design requirement that the data centre user has always stipulated?  That is, the system must meet tight guidelines in controlling temperature, humidity and filtration.

 

There are a number of vendors throughout the world which have been undertaking development of air-side economisers. And the one guiding principle has been to ensure that the data centre air is not compromised by poorly-managed control and by episodic atmospheric events—like wild fluctuations in temperature or humidity, abnormally high particulates from dust storms, or corrosive gases like sulphates—that have every chance of disrupting the cooling cycle in the data centre and literally destroying the equipment that it is meant to protect. How is this attained? By maintaining, as tightly as possible, a closed environment.

 

But, all of a sudden, we have throw-open-the-window merchants who believe they can construct, from scratch, systems to match vendors which have been carrying out years of R&D to refine air-side economisers that meet the stringently controlled environment of the data centre.

 

A number of articles in the mainstream data centre media (articles written by independent sources) present research on this cavalier approach by saying: 'The IT equipment failure rate is low, so why worry? Just replace it, it's cheap.' That's a statement made from one simple perspective. It ain't just about the electronics failing; it's about the actual service the equipment provides to the business and the public. When the service goes down, a lot of people and businesses can be adversely affected.

 

The air-side economiser is a serious cooling contender in the data centre as it has huge benefits in energy saving with subsequent cost savings, not to mention reducing greenhouse gases. What I can't fathom is that data centre users are being lulled into using such a great technological idea without applying the same rigour as they have applied when choosing the closed-cooling systems that are endemic in data centres around the world.

 

 

 

For more information on air-side economisers, here’s a good article by Bill Rumsey.

0 COMMENTS SO FAR...

YOUR NAME:
Your website:
TITLE:
COMMENT:
SECURITY CODE ENTER THE CODE SHOWN ABOVE IN THE BOX BELOW
ADD COMMENT   CANCEL