Adaptive and Fault-Tolerant Service Composition in Peer-to-Peer Systems

Abstract

Abstract. Service-orientation enables dynamic interoperation of distributed services and facilitates seamless service provision or runtime creation of new applications. This dynamic service composition is particularly powerful in peer-to-peer (P2P) systems which offer scalability through self-management and autonomy. However, P2P service composition is nontrivial due to permanent peer churn and lack of central control. Existing approaches reduce composite service initialization to an NPhard path finding problem. Thus, peer failure adaptation is costly and runtime consideration of peer logons or load changes is not practicable. This paper introduces logical peer groups for service composition. They enable runtime composite service reconfiguration including the migration of services to other peers. A prototype implementation is presented and the algorithms are evaluated through both formal and empirical analysis. The evaluation shows that the approach results in significant reduction of computational complexity, improves fault-tolerance and enables adaptation of logons and load changes which has not been possible so far.

Synchronising Service Compositions in Dynamic Ad Hoc Environments

“…This wastes scarce resources for conditional or prematurely-terminated paths and is less flexible towards runtime changes that render selections invalid and cause failure. Dynamic leader election in logical peer groups [13] adapts to dynamically arriving and departing providers but has been proposed only for sequential requests. OSIRIS [5] makes its pre-selection final only during execution, but requires a reliable and globally available synchronisation node to handle parallel service flows.…”

Opportunistic Composition of Sequentially-Connected Services in Mobile Computing Environments

“…However, dynamic service composition in mobile environments is challenging for two reasons: First, service providers may join and leave the network at runtime and their fluctuating presence implies a high failure probability for the composition [2]. Second, a central entity of control is infeasible in such dynamic environments because of the large overhead required to keep global state information up to date.…”

Towards In-network Aggregation for People-Centric Sensing

“…Probing techniques [5,11], primarily used for optimal provider selection, enable service providers to actively commit to a composition rather than allocating them without their confirmation and risking their later refusal. Alternatively, proactive discovery solutions explicitly ask for participation [14]. Both such approaches, however, require additional communication on top of establishing knowledge about available services.…”

“…Binding and releasing services just in time is crucial to maintain good service availability and to mitigate overload situations. Composition approaches that start the execution only after the selection is complete [15] [14][5] [16], allocate all required services even those that do not get executed due to conditional paths or premature termination. They block providers unnecessarily and possibly make them unavailable for other compositions.…”