Bladewatch.com

The traditional way of managing the datacenter is to keep the temperature constant, you set a level that you feel is ideal for the servers, the network infrastructure and whatever storage you’ve got.

In the olden ‘days’ you may have set this value to be say in the 20 degrees range, so that everything remained relatively cool, if there was a air conditioning failure then it would take a while to heat up and therefore time to get everything fixed.

However, as the cost of power goes up, does it not seem as a good time to look at our designed operating guidelines? The average rackmount/blade is designed for 20-40 degrees Celsius, the shelf-life of the server is shortening in real terms, as in where a server used to last up to 8 years, typically in performance computing this has reduced to four (real life years), in which case why not look at your operating conditions?  Would 30/35 degrees really harm your infrastructure?  What are the possible savings from doing this?  Something I’d be interested in looking at.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.