Virtualization adds a layer of complexity to the application performance management equation. So it's critical to have the right tools in place to avoid piling on even more pain.

However, many vendors have been slow to realize that customers aren't looking for standalone virtualization management tools -- rather, they want integrated options that work with their existing management tool investments. Organizations that have dismantled the management silos that looked only at Windows, Cisco, or storage systems won't let virtualization management become another silo.

When working toward a consolidated application performance management, or APM, system that includes virtualization, there are several key areas to keep in mind. Prevention always is the best remedy for performance issues. And the best way to prevent trouble is to ensure that you have configured your physical application correctly for your physical-to-virtual environment. As is the case with physical servers, poor application configuration and unanticipated usage levels on virtual servers can severely degrade production performance.

Virtual machine configuration is another area to consider. Determining how to configure VMs to run best on physical hardware is a challenge, to say the least. Configuration must be based solely on the amount of resources that the application uses. Products like Cirba's Advanced Analytics Engine can help you analyze your existing apps as you plan the migration to virtualization by examining workload patterns and configuration as well as other constraints such as service level and uninterruptible power supply requirements.

Companies with high-availability architectures can use VMware's VMotion technology, which allows an application to be moved to another physical machine across the network. VMware's VirtualCenter live-migration capability can automatically determine when to move apps to different physical servers, helping alleviate downtime if performance becomes an issue, but tracking performance and service-level agreements is more complex, because apps may frequently move among servers.

With virtual environments becoming easier to set up, organizations may find themselves with a plethora of VMs but little knowledge of what's installed or how VMs are configured. Tools that can discover the virtual environment are essential. Traditional discovery tools are topology-focused and execute IP subnet scans, but that won't work for VMs, because a particular host may be in different subnets at different times. In addition, as the VM moves across subnets, a discovery tool must be able to track the state of the VM's performance metrics and assets as it moves.