Conventional data center maintenance programs have been based on a combination of recommendations from original equipment manufacturers (OEM) and best practices that have been adopted by industry groups and trade associations.
Historically, there is very little empirical or statistical engineering science behind them. OEMs believe that more frequent maintenance means lower risk – at least for their equipment. But is this true?
OEM recommendations have become the de-facto maintenance standard because they are required to support warranties, and are bundled into comprehensive service contracts that can easily be renewed for many years. Even after warranties expire, data center operators may be reluctant to risk a service disruption from failure to follow the OEM recommended maintenance.