Why Servers Become Nonproductive
Potentially useful systems remain nonproductive or underemployed for many reasons, including the following. Businesses are unable to predict application-usage profiles, requiring systems to run with little use. New applications replace old ones, leaving the old server running. Old applications retire, leaving the servers running. Business interests change. Business cannot track their IT assets closely, much less the space, power, and cooling they use. More broadly, specific business groups have no idea how their applications and the servers they run on impact the data center environment. This paper covers the many reasons why servers are nonproductive and what can be done about it.
"According to Ken Brill of the Uptime Institute, as many as 30 percent of the servers on the data center floor are not performing useful business work. This statement likely comes as little surprise to many data center operators but they need to be careful about what they do with this information. A data center manager may unwittingly shut down a server that has been sitting idle for 11 1/2 months only to find that the accounting department needs it to run a year-end report. Turning off the wrong server can be a career-limiting move. This white paper examines the problems associated with shutting down servers, as well as possible solutions and their impact." When it comes to managing nonproductive servers, data center operators have become paralyzed because they lack insight into which business applications (if any) are running on which servers and their frequency, or whether or not an application is even being used. Most data center operators live by the Hippocratic Oath of the data center: First do no harm, then take no chances. Faced with the conflicting objectives of maintaining service and reducing energy costs, they automatically turn to the worst offenders by reducing power and cooling