Classification of approaches

 Distributed application scenario   [r]
2#2

In recent years a number of approaches to application performance monitoring have evolved. To alleviate evaluation of specific techniques, the following section introduces a classification gained from a survey of currently existing solutions by applying the requirements mentioned above. Eventually, the most prominent representatives of the different classes - both current products and existing standards - are briefly introduced. Figure  shows a simple model of a distributed application as well as its underlying infrastructure. The measurement of performance parameters can either take place at the application itself or in the infrastructure (both at the network and the systems). When measuring the application itself, approaches that solely focus on the client-side of the application can be distinguished from those taking into account the application as a whole (client- and server-side). This leads to the following four classes of techniques: Monitoring of network traffic, system-level monitoring, client-side application monitoring and application-wide monitoring. The following paragraphs explain these four techniques along with a number of subordinate techniques in greater detail.