Knobs & Gauges

In my experience, the three core services needed for a complete IT solution could be defined as information, resolution, and control. The best way to obtain these services is what I describe as the "instrumentation" of the infrastructure - applying software or hardware "gauges" and "knobs" within the system. In general, the better the instrumentation of the infrastructure, the more granular information, resolution, and control companies receive.

As organizations look at developing information, resolving problems, or implementing control, the depth and breadth of capabilities is based on this instrumentation of the IT infrastructure. By infrastructure, I mean the critical resources that make up an application delivery system, from servers to desktops. The less infrastructure the better, since it costs money to put in place and maintain, which is why instrumentation is so important.

In a perfect world, every element of the infrastructure would provide the proper level of functionality, so that additional software or hardware would not be required. But as we all know, its not a perfect world. Therefore, a critical aspect of performance management is applying instrumentation within the infrastructure. In general, the degree of instrumentation provides a corresponding value in the development of information, the accuracy of resolution, and the granularity of control.

There are a number of key things that should be considered when applying new hardware and software. Within the application infrastructure, there are certain factors that contribute to performance. I view availability as being the most important factor. Next in importance are the services that support availability, which are not only the multiple 9s everyone is striving for, but consistently responsive end-user experiences. In the pursuit of these goals, many organizations have overly provisioned the application infrastructure and dedicated staff for each type of IT resource.

With a wide variety of hardware and software environments throughout the tiers of the application environment, the need to stabilize the IT infrastructure from a configuration and resource-allocation perspective has been the short-term requirement.While there are dozens of companies providing even more products to manage application infrastructures, there are three that come to mind. What becomes clear when looking at these solutions is not whether one is better than the other, but rather the value proposition they are driving to deliver.

Collation’s Confignia has approached the management of the application infrastructure by using what they call ‘sensors’, which initially discover the resources within the infrastructure and then use platform-specific methods to pull information into their system.

Relicore’s Clarity discovers and instruments each server under management, which provides more breadth in the information that can be obtained, tracked, and analyzed. This includes a number of operating system and application resources, as well as the physical inventory of the server hardware, etc.

VIEO’s AIM 1000 instruments not only the servers under management, but the physical networking infrastructure as well. This provides similar configuration management capabilities, but can also manage performance by measuring response times and dynamically tuning infrastructure performance.

The good thing about these three approaches is that we are beginning to see solutions that treat the application infrastructure as a system. Discovering the resources within the infrastructure and understanding the dependencies between them provide significant value in maintaining stability, as well as in speeding up fault isolation when "stuff happens."R. Lynn Nye Jr is president of analyst firm APM Advisors