Monitoring service choreographies from multiple sources

Modern software applications are more and more conceived as distributed service compositions deployed over Grid and Cloud technologies. Choreographies provide abstract specifications of such compositions, by modeling message-based multi-party interactions without assuming any central coordination. To enable the management and dynamic adaptation of choreographies, it is essential to keep track of events and exchanged messages and to monitor the status of the underlying platform, and combine these different levels of information into complex events meaningful at the application level. Towards this goal, we propose a Multi-source Monitoring Framework that we are developing within the EU Project CHOReOS, which can correlate the messages passed at business-service level with observations relative to the infrastructure resources. We present the monitor architecture and illustrate it on a use-case excerpted from the CHOReOS project.