Application end-to-end response time measurement and decomposition

Abstract

A method, computer program product and a program storage device embodying software for measuring of the response time of an application (including distributed applications in a client/server or Internet environment) as perceived by an end-user. One aspect deals with the measurement of components of the response time and relating those components to user response time. In addition, the components of one transaction can be matched (or correlated) to each other even though they are measured on different systems. The generation of events and transactions can be controlled, allowing their creation to occur as close to their point of origin as practical. Both aggregate and detail reporting facilities provide overall performance and availability information as well as exceptions and/or detail transactions including the decomposition of overall availability and performance metrics into smaller measurements representing the contribution made by select transaction components. An interactive reporting facility enables the selection of a level of transaction decomposition desired. This enables the identification of the transaction components that are introducing delays or faults. The system is extensible, enabling the addition of components to the system that extend its measurement and reporting capabilities. In particular, a language has been created to facilitate the definition of the end-to-end business application transactions. Also, select APIs as well as application data structures allow the addition of software and/or hardware modules to extend the system. The system can also adapt to the presence or absence of select streams of events without having to change its mode of operation. Measurement sources that generate events can be dynamically activated and deactivated.