Three Fundamentals for Managing Your Virtualized Infrastructure - Page 2
Current legacy solutions are built as “bottom up” management tools. They are designed to assess the health of the individual elements – servers, routers, switches, databases, applications and storage. They rely on probes and agents to track critical IT elements and are not extensible to the real-time understanding of each and every user’s experience. This approach does not provide the level of understanding and visibility needed to support intercommunications across both physical and virtual domains.
VI administrators need to have a way to measure and monitor the end-user experience from the top down in order to immediately and accurately identify problems for fast resolution. Only then can VI administrators effectively manage the virtualized infrastructure.
Because virtualization relies on a broad and diverse physical infrastructure – with much more interplay between components – using this type of a cohesive management methodology is critical. You need to be able to view any object on your entire infrastructure – physical and virtual – in a way that you can see how each element communicates with the others.
By having the insights into all of the elements as a whole, VI administrators can clearly show where the problem exists. According to Yankee Group, 90 percent of the time spent resolving a problem is just identifying the cause of the problem. Any management solution that helps identify the case of the problem pays dividends quickly.
Having cross-domain visibility is also necessary to identify ownership of a particular performance problem, so that the problem can be readily addressed – rather than passed back and forth between groups. This eliminates a lot of wasted time from the ‘blame game’ by clearly showing the cause of the problem.
Fundamental #2: Visibility Must be Real-Time & Continuous
At the most basic level, IT must be able to view any object on the entire infrastructure – physical and virtual – in a way that shows how each element communicates with the others. But that’s still not enough. Applications and relationships must be tracked in real-time.
Cross silo performance metrics need to be presented and tracked live and continuously to be useful to network, server, storage and application administrators. Administrators need to have a real-time streaming view for each critical object and its interactions.
Legacy management software leaves blind spots between polling intervals and lacks visibility into virtual environments because they are not real-time solutions.
To see performance problems in a virtualized infrastructure you need live, to-the-second video-like visibility into what’s happening between all the different components – storage, applications, networks and servers. If you don’t have live visibility, you can’t capture the dynamic performance changes that occur in an instant. These changes create problems but leave no trace and no way to capture what’s happened.
Having a streaming – to-the-second – view for each application and all of their users and interactions overcome a major roadblock for deploying mission-critical applications.
A DVR type recording also serves this function. For example, showing what server the end user is communicating with, live video (and a recording for later playback) is an invaluable tool to clearly and undeniably target the problem area. So when a line of business owner calls to complain about performance of a key database server, the VI admin can see, for example, that the database server CPU is bogged down by a backup running longer than scheduled and now being hit by workday queries. This reduces finger pointing, proves frequently that virtualization is not to blame and speeds the time to resolution.