Subscribe / Unsubscribe Enewsletters | Login | Register

Pencil Banner

Guest View: VM Zombie hunting – managing the evil VM Sprawl

Peter Chenoweth | Nov. 18, 2013
The flexibility, enhanced productivity and business agility gained by being able to rapidly deploy more resources has also created a new management nightmare in the form of orphaned or ‘Zombie’ VMs

Pre-virtualization

According to a recent global study by Kelton Research, nearly nine out of 10 IT managers believe virtual machine (VM) management is risky without proper planning[1]. While many enterprises have embraced virtualization and its benefits, they now realize it can easily become very much of a double-edged sword. Previously, the complexity and costs of physical infrastructure deployment served as an automatic vetting process for applications. With virtualization, the simplicity of deploying virtualized resources reduces, or in some cases eliminated these barriers, leading to a bigger problem - VM sprawl. 

What causes VM sprawl?


The flexibility, enhanced productivity and business agility gained by being able to rapidly deploy more resources has also created a new management nightmare in the form of orphaned or 'Zombie' VMs. Simply put, Zombie VMs are the forgotten VMs hidden in the shadows of an enterprise's virtual infrastructure, gradually eating away at precious resources; destroying datacentre return-on-investment (ROI), and affecting performance.


VMs are often created to increase resources for short-term projects and then decommissioned and redeployed upon the completion of a project. In reality however, this best practice does not happen as much as it should, especially when there is a lack of clarity on which part of the business owns responsibility for the end-of-life process. Systems/Virtualization administrators often fail to decommission VMs because of there is a lack of clear process and authorisation to do so. This leads them to operate on the premise that it is safer to leave these resources alone and not set a specific end-of-life date and process with the end-user, just in case there is a sudden requirement for them down the line. Developers or engineers, on the other hand, avoid immediate decommissioning because they tend to worry they might need the VM again soon. More often than not, these resources stay unused and are left forgotten. That being said, IT administrators, developers or engineers are not the ones who fork out the money needed to keep these VMs running; hence it is not a top priority for them to make sure any overlooked resources are taken care of promptly.

VM sprawl - Worse than an infection?

 
Almost every virtual environment struggles with VM sprawl to some extent, and while it is a recognized problem in virtualization, it is an issue that has consistently failed to be resolved. With shrinking IT budgets, the 'hoarding' of VMs is costing organizations huge amounts of computing and storage resources that could be much more profitably repurposed to other virtual machines and applications.

 
Although there are a number of specific tools and management software products available in the market addressing this problem, they are primarily reporting tools which only alert administrators and show them where action must be taken. Interestingly, the majority of virtual environments do not deploy any VM management tools due to cost, time or a combination of both.

 

1  2  3  Next Page 

Sign up for Computerworld eNewsletters.