Abstract

ABSTRACT An active XML (AXML) document contains tags representing calls to Web services. Therefore, retrieving its contents consists in materializing its data elements by invoking the embedded service calls in a P2P network. In this process, the result of some service calls can be used as input of other calls. Also, usually several peers provide each requested Web service, and peers can collaborate to invoke these services. This often implies a huge search space of many equivalent materialization alternatives, each with different performance. In this paper, we model AXML documents from a workflow perspective and propose a dynamic cost-based optimization strategy to efficiently materialize them, considering the volatility of a typical P2P scenario. Our strategy enables the optimizer, called XCraft, to get more up-todate information on the status of the peers, and to deliver partial results earlier. Based on a service-oriented algebra of plan operators, we exploit P2P collaboration to delegate both execution and optimization control. Our tests with an XCraft prototype show important performance gains w.r.t. a centralized approach, whilst the optimizer also achieved to drastically reduce the size of the search space.