The functional constraints of service matchingand performance requirements (i.e., quality of service, QoS)of service selection are separated distinctly under thetraditional service-oriented architecture (SOA). Hence theservice matching and service selection may be performed inturn and independently. And the complexities of theimplementation of SOA are simplified also. In fact, thefunctional constraints and QoS are stretching andcompromising frequently. And service providers withsimilar functions would come into being a merged-servicewith more powerful function and higher performance bycollaboration with each other. This paper proposes anextended QoS (exQoS) system that contains some servicefunctional constraints and traditional performanceindicators. Then a service merging enabled SOA (meSOA) isestablished based on the exQoS, and the meSOA may beused to support service merging across enterprises.Furthermore, the problem of service matching and selectionbased on merged-service is formulated as a nonlinearplanning (NLP) problem, and its corresponding solvingalgorithm is proposed also. At last, this NLP method andtraditional method are compared with a numerical valueexperiment, which verified its correction and the superiorityof this method over the traditional one.
展开▼