subject: If A Data Centre Overheats, Hired Air Con Can Save The Day [print this page] Equipment failures at data centres (also sometimes called server rooms) often make the news, particularly if hundreds or even thousands of people world-wide have been affected by a data centre breakdown.
Not all breakdowns are caused by equipment overheating due to excessive temperatures within a data centre room or dedicated building. But when a built-in air con system develops a fault, or simply does not have the capacity to cope with extreme heat, this can create chaos that can ultimately erode a companys profits as well as its reputation.
The good news for data centre managers is that hired air con is quick and easy to arrange through a good air con hire company, and is surprisingly affordable.
Why Data Centres Sometimes Overheat
Data centres can be large or small, depending upon their required capacity. Regardless of their size, what they will all contain are server computers: huge units that serve desktop computer users needs on a network. Several servers can be housed within a data centre, where they are secured in racks.
Related equipment will also be found in a data centre, where a team of engineers will work under the direction of an Operations Manager. For security reasons, and to prevent the accidental switching off of the built-in air con system, access to the data centre will be strictly controlled.
It is the combination of 1) the accumulative heat created by servers, and 2) high summer temperatures that usually causes system performance problems and sometimes complete
breakdowns at data centres. In summer a data centre can be particularly vulnerable to failure as on average the centres temperature will be 10C higher than the rest of the building in which it is located, meaning equipment can rapidly overheat. When the temperature reaches dangerous levels, server malfunctions are inevitable.
Examples of Data Centre Failures World-Wide
Every year, data centre failures occur all over the world. Here are just a few examples that have made the news:
In April 2007, the temperature between some server aisles at Facebooks data centre in the
U.S. reached 43C. The Plexiglas sides of some server racks warped from the heat, according to the DataCentreKnowledge website.
In June 2009, due to a cooling failure, several servers overheated and crashed at Level 3
Communications data centre in London. Although no system downtime occurred there,
this was not the case for the London-based music service Last.fm, whose data centre also
overheated in the sweltering summer temperatures that year, resulting in a loss of service for approximately five hours, states Techworld News.
In December 2009, due to power problems at a data centre in northern Virginia in the U.S.,
Amazon website customers experienced a loss of connectivity and so were unable to purchase goods. It took several hours to get the Amazon site back online and fully functional again. (A lightning strike in Dublin, Ireland, caused power outages at the major cloud computing data hubs for Amazon in August 2011)
In March 2010, Wikipedia (one of the worlds busiest websites) had to temporarily go offline
after a cooling problem at their European data centre in Amsterdam led to a heat condition that caused a server shutdown. Many of the 50 servers housed there switched themselves off when temperatures reached hazardous levels.
In August 2011, a major Primary Care UK software provider profusely apologised and launched a formal investigation after its data centre overheated, causing server computers to fail.
In October 2011, due to a data centre problem that lasted six hours, millions of Blackberry
users across Europe were unable to access their e-mail, instant messages, or the internet. The problems in the EMEA region (Europe, Middle East and Africa) were traced to a failure at a data centre in Slough, as reported by DataCentreKnowledge.
Contingency Planning is the Key
Repairing data centre equipment can be a long and difficult process, involving huge cost. Fortunately, temperature control and cooling is quite easy to get right, especially if a data centre Operations Manager forges a link with an air con hire specialist that:
- Has a proven track-record of providing emergency cooling solutions at data centres (where breakdowns have occurred)
- Has fully-stocked equipment depots strategically located around the UK
- Performs free site surveys (as part of contingency planning formulation)
- Offers a full 24-hour 365-day nationwide service meaning they can respond quickly in an emergency
All this will give a data centre Operations Manager peace of mind, knowing that an air con specialist will be on site quickly, should a data centre failure occur. Because a free site survey has already been performed by the specialist, they will know which mobile air con units would be most appropriate for the centre. The temperature will be quickly lowered, with minimal or no disruption to normal server operations.
