Suppose there are 4,000 servers in a data center for Company ABC. These servers may have been installed at different times, and not together, in the same month or week.
A typical server has a maintenance cost of nearly $2,000 a year. On average, only about 75% of the servers are active as they have the task of hosting the most active applications. But since the entire set is up and running all the time, Company ABC is burdened with an extra $2,000,000 every single year for virtually doing nothing. This is just an example. Just imagine a company like Facebook having servers doing nothing part of the time - the wastage would be exorbitant.
This happens in every corporate environment. In order to tackle the spiraling maintenance costs, proper organization of active and zombie servers is necessary. The so called inactive or zombie servers need to be carefully sorted out and treated separately to reduce such huge losses. Careful documentation of each server and the type of applications they run can help a great deal in switching them off when the usage of the app ceases.
Of course, this wastage is compounded when an enterprise operates in different continents and need to maintain local or country based data centers for optimum efficiency. Either way, it is imperative to pay attention to the activity and inactivity of each server is imperative to getting rid of as much data center waste as possible.