Mar 11, 2011 (07:03 PM EST)
APM And Virtualization
Read the Original Article at InformationWeek
But while virtualization can help companies reduce sprawl and bring IT infrastructure costs under control, the gains could be limited unless they effectively manage application performance. Virtualization's benefits come with some significant challenges from an APM perspective. Each application instance typically runs in its own virtual machine, which can complicate monitoring and troubleshooting; APM tools must adjust to the presence of a virtualization layer between applications and their physical platforms.
Inside The Hypervisor
Almost two-thirds of companies are extending the APM tools they've been using for nonvirtualized environments into the virtualized realm, according to our survey. The question that companies need to answer is whether the tools are adequate for virtualized applications.
The hypervisor is a key data source for APM tools. It manages the sharing of resources by multiple operating systems running on the same physical server, as well as the spawning of VMs, scheduling and prioritization, and other tasks. The hypervisor's prioritization decisions can affect application performance.
Given this, IT needs to have visibility into the hypervisor's behavior. Until recently, it has been difficult for APM tools to get access to this data. Now, however, most virtualization platform vendors provide management systems that oversee hypervisor activity and serve as the single source of information about them. VMware has set the pace with its vCenter Server, but Citrix, Microsoft, and others have developed similar systems. APM tools can use APIs to tap into these hypervisor management systems.
APM tools use APIs to detect, diagnose, and report on how problems in the hypervisor are affecting applications contained in VMs. An increasing number of APM tool vendors also are using APIs to supply data for agents positioned inside VMs and in the applications themselves to monitor transactions. Examples include BlueStripe, Compuware, DynaTrace, OpTier and Quest Software.
Access to the data is also important for providing visibility into dynamic load balancing. A variety of tools, including virtual application delivery controllers, are able to monitor and interpret this data as workloads change on servers due to the hypervisor's provisioning adjustments.
In addition to third-party tools, the virtualization platform vendors also have products that play a role in managing this aspect of resource utilization. Examples include VMware's Distributed Resource Scheduler and Microsoft's System Center Virtual Machine Manager. These tools can balance VM loads and monitor and allocate resources according to priority rules and thresholds.