REDUCE–REUSE–RECYCLE: How to Better Utilize Your Virtual Infrastructure

Virtualization has improved hardware efficiency via server consolidation, and has also dramatically reduced the time required to bring a new server online from, weeks or months down to hours or even minutes.  This ability to bring machines online quickly provides companies with improved business agility.  However, as with most improvements, there are also some new issues created which will reduce a virtual infrastructure's overall efficiency over time if not addressed.

Virtual machine sprawl is one of the biggest concerns facing many companies that have deployed desktop or server virtualization.  The ability to quickly create virtual machines without the same disciplines and controls that are in place in the physical world results in machines being provisioned unnecessarily without proper justifications and approvals, machines being over provisioned (too much CPU, memory or disk),  and machines consuming resources well after they are required.

To address this need, there have been a number of reporting products created to help track and identify virtual machine sprawl.  The problem with report only solutions is that they provide a partial solution.  They only help identify the problem and track it if it is getting better or worse.  They do nothing about preventing the problem from occurring or correcting it once it has occurred.  This article looks at how virtual machine lifecycle management tools can help decrease virtual machine sprawl by automating the process of reducing, recycling, and reusing compute resources in a shared physical infrastructure.

The Roots of Sprawl

If you look back at the IT industry, virtual machines are not the first compute resource to have sprawl problems.  The general rule is that the easier it is to create something, the more of them you get, and the harder it is to clean up.  Emails, files, databases, and storage are just some examples of information technology that also have sprawl issues.  Reporting is a key component to identifying the problem as well as the scope.  However, when it comes to fixing the problem, reporting is just part of the overall solution.  With any technology that is susceptible to sprawl, policies and governance along with automation to help clean up and recycle are necessary components to achieving efficient resource usage.

Take email as an example of sprawl.  Most people can easily find out how many emails they have, how much space they have consumed, and even which emails are consuming the most space.  Those reports, however, do little to prevent email sprawl.  The only way administrators control email resource consumption is by  implementing policies that control the amount of capacity each user could consume with their emails and how large and what type of attachments were permissible.  Additional policies that automate the archival of old emails or the deduplication of common information further optimize email resource utilization.

Compute virtualization is no different than these other IT resources that have caused sprawl issues in the past.   There are two primary issues why virtual compute infrastructures have more sprawl issues than their physical counterparts: