The Complexity Factor

For IT, the issues aren't just around virtualization as an evolving technology. The issues also center on people, processes and tools. In the physical world, IT was divided into areas of expertise--servers, applications, networks and security--and could be called upon to support users in any one of their domains. Virtualization, because of its evolving nature, has created a self-service mentality where individuals want to claim new systems and provision them on their own. Templates are being created and emailed to a colleague and VMs are being moved to different storage systems. This in turn has created a blind spot for IT. Take for example an organization with several VMs on a single piece of hardware that are interfacing to a storage system. Alerts from the storage that the device was about to fail were being stored in a log, but the administrator didn't know the log existed and the entire infrastructure failed after 30-days.

Provisioning, dependency mapping, patching and auditing are all new dynamics in the virtual world. 

Virtualization Management Tools

Today the management infrastructure for virtualization isn't as mature as that of the physical world. Existing tools to manage and monitor virtual environments are satisfactory for today, but as virtualization moves beyond server consolidation and into production environments, configuration management will become even more critical.

So what tools exist and what are some best practices for configuration management? One of the biggest arguments / debates tend to come from the VI admin and the application owner on how much available resources are needed for the virtual machine.  Application owners want as many resources as possible dedicated to their VMs. However, the VI admin typically concentrates on supplying the amount of resources the VM will truly need. The Virtual Infrastructure has become the new network group or scapegoat in that any problems, not matter what they are, automatically become a virtualization issues.

So what can be done about this and what are some best practices for managing virtual infrastructures? There are many best practices that can be employed to help VI engineers and admins make the most of their virtual infrastructures.

Building a Virtualization-Aware Management Infrastructure

A lofty goal, but one that is attainable is to drive down the cost per VM by running more VMs per physical server without sacrificing availability and performance, while keeping administrative costs low. Management tools that assist with these needs are available today.

To achieve both cost and performance goals, IT professionals should select management tools that can:

• Aggregate information - management tools must pull together information from multiple virtual and physical machines to present a complete picture of issues that could affect performance or availability. Many legacy tools can run their agents on individual VMs but cannot understand the whole picture, as events affecting a VM may exist outside its field of view. Virtual infrastructure management tools need to pull data from the physical infrastructure and well as VMs. Administrators get a holistic view of infrastructure health, without having to visit multiple point products to collect all the information they need.

• Consolidate management tasks - As noted earlier, the number of hypervisors in typical data centers will increase as vendors offer increasingly interesting alternatives. Over the next year or two, virtual infrastructure management tools will need to integrate with the top virtualization platforms.

• Make automated and intelligent decisions - Managing a large numbers of VMs requires basic automation for frequent tasks like patching or provisioning. However, in order to run at maximum utilization, management tools must quickly detect problems and correct them - down to the root cause.  

It's Not Just About the Tools

As you upgrade your virtualization management capabilities, keep in mind that achieving ROI is more than just technology. People and processes need to be addressed as well.  Just as the technology surrounding virtualization changes, so do best practice policies.  For example, make sure that your VI team attends change management meetings and looks at their charge back policies.

Virtualization means not only new technology, but new ways of managing technology. According to Forrester, "Because server virtualization consolidates server, storage, and network resources into a single box, firms are discovering that some slight changes are necessary.

For example, it can be impractical to train all administrators on server virtualization."  As a result, having a dedicated group of virtual infrastructure administrators will allow other server administrators to focus on running the operating systems and applications inside the VMs.

Tripwire , Virtualization Management