Additionally, many organizations have not reached the point where latency, rather than utilization, is the issue. While utilization is a small part of a bigger picture, latency and response time tells the whole story. In fact, organizations utilizing proper virtual machine management strategies spend half the time dealing with unexpected issues as compared to those who do not, according to the study by Kelton Research. Technical details, such as knowing appropriate VM density, the way VMs interact, and determining the best configuration based on those factors, lead to significant efficiency gains, which after all is the end goal of any IT or business unit.
The elixir vitae for VM sprawl
To help organizations trying to eliminate waste and reduce resource consumption while driving higher utilization efficiencies in their virtual computing infrastructure, there are three basic rules to remember about VM management:
Gaining better control is considerably easier than trying to identify over-provisioned machines at a later date. By controlling requests for new virtual machines and ensuring each request receives appropriate review and justification, IT administrators can delegate the provisioning process downstream to resource users. With appropriate governance policies in place, IT administrators can also control how the VM is built and the amount of resources it can consume.
A core doctrine of virtualization is improved efficiency through better resource utilization. The overall efficiency of virtual infrastructure is highly impacted by the degree to which resources can be reused. With proper administrative control, policies or parameters can be put in place to dictate the lifecycle of a VM, while process automation ensures the easy management of VMs. Inactive VMs containing valuable data can be backed up, protected and archived, to be revived later if needed. Think of it as VM 'reincarnation'.
Identification and collection of unused resources is often a time-consuming and menial chore. Hence, it is no surprise the process is performed less than it really should be. By identifying stranded capacity, which is when one of the three most critical resources (CPU, memory or storage) is completely consumed while the rest of the resources are abundant, IT administrators are able to relieve the bottleneck. This can be accomplished by adding capacity to that particular resource. Another option is to reallocate stranded capacity to other hosts, improving the overall utilization of the virtual computing infrastructure. Taking into account a list of criteria to identify potentially inactive VMs is also very helpful in the decommissioning and reclamation of resources which are no longer needed.
Data can be pruned over time, and while it might not be the easiest job, somebody has to do it. While most organizations do not deploy VM management tools in their virtual environments, backup and recovery tools are de rigour to ensure business continuity. The addition of automated, customized policy-driven processes to the backup and recovery tools already in place on a singular platform makes it simple for organizations to eliminate VM sprawl. That way, organizations can concentrate on growing their business, and not worry about finding the best way to deal with Zombie VMs.
Sign up for CIO Asia eNewsletters.