A recent survey conducted by Network World revealed that most IT managers were unable to measure end-user experience with their traditional network monitoring software tools. Over 50 percent of the survey respondents identified page response time, server query response time and TCP transaction response time (key measures of end-user experience) as being important, yet were not able to measure these metrics with their existing management tools. The survey highlighted a need for IT and network management software that is able to monitor the performance of IT from a user perspective (e.g. end-user page response time), as well as monitor the performance of the various underlying network, server and application components that make up the layers of infrastructure that enable delivery of services.
Although there are specialist solutions that support end-user experience monitoring, these tools are generally not pre-integrated with management tools that monitor the health of the underlying IT infrastructure. Having the linked ‘top-down’ and ‘bottom-up’ views and integrated capability within one IT monitoring system allows tracking service performance and user experience metrics, and then if problems are detected, the solutions facilitate drilling down to view and analyze the technical performance metrics for the various enabling components (e.g. CPU utilization of the application server). This capability allows rapid and context-specific identification of potential causes of degradation of end-user experience. Having unified, correlated, status views allows the IT team to not only better assure the real-time user experience, but also conduct detailed analysis on areas of performance issues and bottlenecks in the underlying IT infrastructure.
Organizations that are in the midst of exploring new network monitoring software solutions can look for the following types of capabilities to get an integrated view of performance. Does the solution monitor metrics, such as response time, for complete multi-step end-user transactions? Ideally, any number of multi-step test transactions should be definable, where these tests can be monitored alongside the other device or server specific tests to generate alarms when thresholds are violated. As part of scripting a transaction step, the user should be able to select specific frames and links for navigating through a particular path for testing purposes. Additionally, secure pages should be accessible by providing the relevant authentication credentials. As part of the scripting process, when the user clicks through to the next step, the software needs to be capable of performing basic validation to ensure that the transaction being scripted can indeed be executed without application access errors.
Combining transaction monitoring and infrastructure monitoring in one system, and then taking this one step further by mapping services to the relevant top-down and bottom-up metrics (see example of service monitoring solution at http://tiny.cc/mpqxn ), allows organizations to monitor service performance from both technical and end-user perspectives. As the overall IT infrastructure becomes more dynamic and complex with adoption of new technologies such as virtualization and cloud, the ability to unify and tie infrastructure monitoring with end-user experience monitoring will allow organizations to better assure overall business performance and customer satisfaction.