Achieving Effective Application Performance in a Virtualized World

Aimed at enhancing data center operations, IT departments are steadily moving toward virtualization technologies because they open up many possibilities in terms of flexible response to peak demands, agility and productivity through automation, as well as cost and computing efficiencies in the areas of hardware, power and cooling.

But virtualization also adds another layer to an already complex IT infrastructure. If virtualization slows down important enterprise apps, or makes it more difficult for IT to find and fix problems, the damage can result in lost productivity of employees, lost revenue and lower customer satisfaction.

Even if your data center monitoring tools indicate strong infrastructure performance, your end users may be having an entirely different experience, and it’s the perspective of the end user, not your infrastructure that matters.

Troubleshooting Virtualized Applications

Troubleshooting virtualized applications is a big challenge for even the most seasoned IT professional. Where do you look and what do you look for? Data center monitoring tools do not track the end-user experience and are not completely effective in a virtualized environment.

To track down and isolate application performance issues, IT organizations have tried using component-level monitoring tools and tools from virtualization vendors. Unfortunately, neither of these tools sets is completely effective in a virtualized environment.

Component-based monitoring tools take a bottom-up approach, starting with the low-level infrastructure. They may give some visibility into the performance of the physical and virtual infrastructure, but none into the applications running in the virtual hosts. Many metrics collected by these lower-level tools – such as CPU usage, network load and disk I/O – are not applicable in a virtualized environment. Furthermore, component-based monitoring tools simply don’t convey how these factors relate to the end-user experience, which is what truly matters.

Tools from virtualization vendors can monitor the virtual environment, but many applications consist of virtual and physical infrastructure that bring together components from both modern and legacy applications across the data center and external third parties. Virtualization instrumentation and hypervisor health, while necessary, only provides a partial view of application performance. In addition, metrics like response time and transaction time collected from within a virtualized environment do not provide a true view of end-to-end application performance for your end users (that is, unless your end users live in your data center, which is highly unlikely).

Lacking the right levels of visibility and troubleshooting information, IT organizations are not able to efficiently address performance problems. Many businesses have dozens of people committed to a war room involved in the development, delivery and support of a virtualized application. However, this approach can tie up a huge number of resources and often yields little besides excessive finger pointing, while end users and the business suffer.