January 2014 VCDX56 sponsor blog by SolarWinds

The new element, i announced in this blog post, where my blog sponsors get the chance to publish their own blog post starts now. As i mentioned, there will be one slot open for blog sponsors per month and the January one is covered by SolarWinds.

Application-aware Virtualization Management

A virtualization infrastructure, when not monitored properly, could turn into a complex maze of performance issues. There are so many dependent parts in a virtualization setup such as the host, cluster, storage and other shared resources which could cause poor VM performance if you do not have enough visibility into their health and performance. As virtualization administrators, we often tend to only look at the virtual machines (VM) and the possible bottlenecks arising around them. Without any doubt, this is needed to troubleshoot VM performance issues, and that’s what a virtual admin or data center admin is supposed to do; but there’s more to a VM than just the VM itself. Any issue with the VM or underlying resources in a virtual environment will most certainly impact the applications running on them: causing them to either go down completely or have poor performance. This is a bigger problem as it impacts the end-users more directly and, in turn, your organization’s business services. Application performance and VM performance are not two different states of affair. They are two sides of the same coin – and the coin being continuous and uninterrupted IT service which is only possible if all the pieces of the virtualization puzzle are addressed including the applications and application servers.

Understanding Virtualization Management

Virtualization management is the process of monitoring various aspects of the virtual environments including VM, hosts, clusters and datastores to ensure they are all performing as expected. Virtualization management doesn’t stop here. Once you have all your performance metrics in check, you need to ensure if there are any capacity bottlenecks and resource contention issues. Many times VMs fail when they are contending for more resources than they are allocated.

Then, there’s the celebrated VM sprawl, which when left unchecked, could develop into a virtual avalanche leaving behind many locked in resources associated with unused VMs.

All these aspects of virtualization management do not only require VM admins to monitor the current performance of the VMs, but help to right-size the entire virtual environment and forecast host resource procurement and budget planning.

Where Should You Have Application Awareness?

All the above factors are good to maintain healthy and consistent VM performance. But do you have enough visibility to ensure business applications are running properly on the VMs? Application awareness is key in multiple facets of server and application virtualization. When you are deciding to move to a virtualized infrastructure to run your apps, you need to know what apps are running, how many are business critical, and then decide what would make more sense to virtualize.

  • It is possible that once you run apps on a VM, there could be performance issues because of any chinks in the application’s compatibility to run smoothly on a VM.
  • It is also possible that the application tends to run slower on a VM compared to running on a physical server.
  • What about the number of applications you run on a VM? Is the VM aware of its resource capacity and can it withstand the application load?
  • From the monitoring perspective, without proper application performance awareness, it could always be taken for granted that it’s a VM issue that’s impacting the application – which is not always true: the app itself could have some latency. This can only be identified if there’s application awareness.
  • Then, there’s the storage case. An application running on a VM has its storage shared with other VMs causing another resource bottleneck in either storage I/O or disk capacity. VMs start to vie for resource, and those VMs that do not have enough resources to run the app, eventually succumb to application availability and performance issues.

When there’s an application issue reported, especially those that are running on VMs, it’s always tricky to base the reasoning on the application itself or the virtualization entities. All this makes it significantly important for IT teams to have contiguous visibility from the application, to the VM, to the host, and down to the storage.

While virtualization management is performed to ensure VMs do not create application issues, it is also important to ensure from the application-side that the application does run well on the VM. There can be so many false positive VM alerts avoided when we have proper understanding of the application performance metrics in relation to the VM performance metrics, and when you are able to map which host is running which VMs, and which VMs are running which apps.

Benefits of Application-aware Virtualization Management

Having contextual visibility across the application stack allows for:

  • Quicker problem diagnosis – whether on the app-side or VM-side
  • Faster troubleshooting and issue resolution
  • Better understanding into how applications are performing and a constant tab on their resource load on the VM
  • More efficient application and server virtualization capacity planning
  • Simpler and centralized monitoring

There are tools that provide virtualization management capabilities, and there are application monitoring tools that offer deeper application insights. An integration between the two establish a single-pane-of-glass control center to look through and across the application stack at just one go. For e.g., to monitor a SharePoint application accessed by a developer in China, run by an ESXi host in Germany, which is monitored by the IT team in Canada, it is possible to demarcate the app issue and the VM issue, and make your troubleshooting much quicker, only if you make your virtualization management application-aware!

About the blog post author: Vinod Mohan is a Product Marketing Specialist Team Manager at SolarWinds who has technical expertise in IT management areas spanning virtualization, storage, applications, and network management.