2010 will be the year that many enterprises confront two very important changes to how they will use server virtualization. The first change is that as VMware vSphere has proven its maturity, performance and scalability enterprises will increasingly put business critical tier applications, at least in part on virtualized platforms. The second change is that at the same time, these very same enterprises will start to evaluate virtualization platforms from other vendors, in particular Hyper-V from Microsoft.
Since these two evaluations will be occurring concurrently, it will only be logical and prudent for enterprises to ask if the very same tools that would be used to ensure the performance of tier one applications on a VMware platform are also available on the virtualization platforms from Microsoft, Citrix and Red Hat. The good news is that with the exception of VMware AppSpeed and a few VMware specific features in a few of the products, these products are largely much more focused upon the applications than upon the specific nature of the virtualization platform, and are therefore largely platform agnostic.
The diagram below identifies four categories of Applications Performance Management Solutions:
- Application Performance Management which focuses upon primarily the measurement of response times between the tiers of the applications. BlueStripe is the only APM solution that can automatically calculate the hop by hop response time for any application across any physical or virtual infrastructure. Quest and OPNET focus on web-based applications but these products also contain extensive diagnostics into the J2EE and .Net applications stacks which helps developers find issues in code. VMware AppSpeed also focuses on web applications with database back ends, but is unique to the VMware vSphere platform.
- Transaction Performance Management which focuses upon the response time and tracking of individual transactions from end-to-end across the applications system. Optier pioneered this category and has the broadest support in terms of applications, operating systems and middleware platforms. dynaTrace and Quest focus upon web-based applications and also contain in depth diagnostics of code in the J2EE and .Net application stacks.
- End User Experience Management which focuses upon application response time from the perspective of the actual end user PC. These solutions are unique in they are the only solutions that actually measure response time all of the way to the “pane of glass” and include client render time (the time to paint the screen in the browser or the fat client application) in their calculations.
- Virtual Desktop Performance Management and Assessment. Liquidware Labs pioneered this category of solutions that focus upon helping VARs and customers characterize their physical end user environments, enabling an intelligent migration of the correct candidates from a physical desktop to a virtual desktop. These very same solutions also can then monitor the resulting production virtual desktop environment for applications performance. This is a critical capability for an virtual desktop deployment since as soon as a user’s desktop moves from “their” PC to “IT’s data center” IT becomes responsible for its performance.
- Performance Management for Cloud Hosted Applications. New Relic has pioneered this category with a unique hosted solution that focuses on Ruby-on-Rails and Java applications but is completely independent of the underlying physical or virtual infrastructure.
It is important to note that many of these solutions can also monitor the performance of the application on physical infrastructure. BlueStripe, Quest, OPNET, dynaTrace, Optier, Aternity, Knoa, eG, LakeSide, and New Relic all have substantial businesses monitoring applications performance for applications residing on physical servers and desktops. Liquidware is usually initially deployed on physical desktops, with the obvious intent of following those desktop into a virtual environment.
The exception on both the question of support for other virtualization platforms and support for physical platforms is VMware AppSpeed. This is because AppSpeed is a virtual appliance that relies upon the mirror or spanned port on the VMware vSwitch, and because its user interface is a tightly integrated vCenter plugin. Some of the other products also contain VMware specific features. For example, dynaTrace relies upon a VMware specific API to rationalize response time measurements in response to the timekeeping issue. OPNET (like AppSpeed) has a virtual appliance that attaches to the virtual mirror port on the VMware vSwitch (which would obviously not work on the other virtualization platforms), but also has physical version of this appliance that works when applications are deployed on other virtualization platforms or on physical servers.
In summary the Applications Performance Management industry is ready and waiting for both tier 1 applications on a vSphere platform, and those very same applications on other virtualization platforms. Enterprises should take a strong look at these solutions since these solutions can handle both virtual and physical deployments and can handle deployments across virtualization platforms. Since tier 1 applications are likely not to be virtualized in their entirety in one fell swoop a solution that can monitor across physical infrastructure, and one or more virtualization platforms should strongly be considered.
For more information on these solutions as well as on the separate category of infrastructure performance management, please download the white paper below.
TVP: This paper discusses how virtualization and cloud computing create new requirements for performance monitoring and capacity management, and how those requirements must be met in next generation solutions. Products from leading vendors are compared against these new requirements.