David Messina, VP Marketing and Product Management, Xangati
Today marks the last in a series of tips for helping our existing and potential customers effectively manage their VDI deployments. If you are at the point that you can leverage Tip #5, then you are in a very enviable position—it means that you have progressed out of the pilot phase and been given license to expand your virtual desktop implementation on a much broader scale. However, now the challenge at hand remains the fact that scaling up an implementation of this level of complexity does not always come with economies of scale—which is precisely why [Tip #5] Scaling Deployments Across the Globe Requires Proactive Service Assurance.
No matter how good your operational support infrastructure is in terms of problem identification, you will have to move out of reactive mode if you want to keep your head (and the head of your team) above water. That means, relying on solutions that get you in front of problems before the ripple effect adversely impacts the end-user. There are just so many moving parts within the virtual desktop ecosystem—including the elements that interact with it—that trying to instrument an alerting framework based on classic manual thresholding could be a dizzying experience, to say the least.
The challenge with threshold models comes in six different dimensions:
- Which long list of elements do you need to track—especially when new things are being added all the time to your infrastructure?
- How do you decide what is the right threshold level?
- If your thresholds are static, how do you keep up with the evolution of your infrastructure?
- How do you account for different periods of the day when activity levels may vary significantly (e.g. back-ups only occur between 2-4am)?
- What do you do in scenarios where hypo-activity is a critical indicator of a problem (e.g. when a specific server that should be handling 30 concurrent users is only supporting three)?
- Given the need to be aware of issues ASAP, is it sufficient to be warned two or three polling intervals/10-45 minutes into a problem state?
Given the challenges noted above, it is clear that the traditional management systems you have been leveraging will not migrate smoothly into the dynamic world of VDI. The Xangati virtual management solution, on the other hand, was specifically designed for these types of dynamic environments. Let’s see how it tackles each of the issues noted above:
- “Out of the box” it automatically begins to track the behavior of every device and application—both virtual and physical—on the infrastructure .
- The system is the one that does the learning of what Xangati calls “profiles” instead of putting the burden on the administrator.
- The system also continues to re-assess and modify the profile over time to reflect any gradual changes that may have occurred.
- The behavioral profiles will have a periodicity to them, if the activity of the element is time-based.
- The solution accounts for not only hyper-activity but also hypo-activity—because, in many ways, the absence of activity is more critical in terms of something that will affect the productivity of an end-user.
- Given the Xangati solution is looking at all communications live also means that it is continuously comparing that activity versus the profiles; and therefore, if there is an issue with an element, the alerting mechanisms will be real-time. Lastly, in addition to the traditional alerts sent out, the system will also initiate a DVR recording at the precise time that an issue occurs so there is full detail of what was going on at the time of a problem.
When put into practice, the proactive capabilities noted above, allow for an extremely scalable framework by which to confidently expand your VDI solution.