QoS Analysis in Heterogeneous Choreography Interactions

With an increasing number of services and devices interacting in a decentralized manner,

choreographies

are an active area of investigation. The heterogeneous nature of interacting systems leads to choreographies that may not only include conventional services, but also sensor-actuator networks, databases and service feeds. Their middleware behavior within choreographies is captured through abstract interaction paradigms such as

client-service

,

publish-subscribe

and

tuple space

. In this paper, we study these heterogeneous interaction paradigms, connected through an

eXtensible Service Bus

proposed in the

CHOReOS

project. As the functioning of such choreographies is dependent on the Quality of Service (QoS) performance of participating entities, an intricate analysis of interaction paradigms and their effect on QoS metrics is needed. We study the composition of QoS metrics in heterogeneous choreographies, and the subsequent tradeoffs. This produces interesting insights such as selection of a particular system and its middleware during design time or end-to-end QoS expectation/guarantees during runtime. Non-parametric hypothesis tests are applied to systems, where QoS dependent services may be replaced at runtime to prevent deterioration in performance.