In theory, IT service management should contain sprawl, limiting or preventing the spread of underutilised IT assets.
Any organisation that prides itself on being lean will recognise the contradiction in multiplying the number of half-used servers that then take up twice the space and power without twice the results.
So why do some large corporations still manage to run huge server farms for little effect and significantly increased risk of failure because all those half-used (if that) assets are also interlinked with each other?
A recent article by The Register singles out a British Airways service failure, emphasising the high number of servers required for what appears to be relatively small tasks.
One possible explanation is that servers were added as new functionality was required, but that the relatively high reliability per new server meant the IT department could, in a sense, “set it and forget it”.
However, internal inter-dependencies between systems create new risk, making any “set it and forget it” approach unsuitable for the server population viewed holistically.
Cloud computing and the ease of spinning up virtual machines does not make server sprawl any easier to control. In fact, it encourages it.
Sure, it is as easy to spin down VMs as to spin them up, but who remembers to do that? Some organisations find out the hard way when their cloud service provider sends them the bill, including the charges for all the virtual servers that were turned on without being turned off.
Others are somewhat smarter and use IT service management to check that the total amount of capacity used does not unreasonably exceed the total amount required. Besides keeping servers in check, the same approach helps redistribute desktop PCs, laptops, and software licenses, rather than taking the easy way out of just ordering more.
The right ITSM then not only saves money, but also improves security, while still meeting the needs of stakeholders, whether internal or external. As they say, “what’s not to like!”