Adaptive SLA Monitoring of Service Choreographies Enacted on the Cloud

Abstract : The deployment and the execution of applications on dynamic Cloud infrastructures introduces new requirements of adaptability with respect to monitoring. Specifically, the governance of service choreographies enacted over Cloud-based solutions relies on the observation and analysis of events happening at different abstraction layers. Adaptability requirements are even more evident when monitoring deals with Service Level Agreements (SLA) established among the choreography participants. In fact, as the Cloud paradigm offers on-demand solutions as a service, often monitoring rules cannot be completely defined off-line. Thus also the monitoring infrastructure must keep track of the continuous evolution of the underlying environment, and adapt itself accordingly. This paper proposes an adaptive multi-source monitoring architecture that can synthesize on-the-fly SLA monitoring rules following the evolution of the Cloud infrastructure. We demonstrate the idea on a case study and discuss limitations as well as planned further advancements.