Composition of Autonomous Services with Distributed Data Flows and Computations

FICAS, an experimental Flow-based Infrastructure for Composing Autonomous Services, supports a service-composition paradigm that integrates loosely-coupled software components. The software components – processes managed by independent service providers – are autonomous services. FICAS employs a distributed data flow that differs markedly from centralized processing adopted by many current service integration frameworks, such as CORBA, J2EE and SOAP. Distributed data flows support direct data exchanges among autonomous services, and thus avoid many performance bottlenecks attending centralized processing. The performance and flexibility of FICAS are further improved by adopting active mediation, which distributes computations within the service framework, and reduces the amount of data traffic significantly by moving computations closer to the data. Several varieties of project management and scheduling software have been aggregated using the FICAS prototype. This demonstration establishes that distributed data flow, combining with active mediation, is more effective and efficient than centralized processing when integrating large engineering software services. 1 Ph.D. Candidate, Department of Electrical Engineering, Stanford University, Stanford, CA 94305. E-mail: [email protected] 2 Research Associate, Department of Civil and Environmental Engineering, Stanford University, Stanford, CA 94305. E-mail: [email protected] 3 Professor, Department of Civil and Environmental Engineering, Stanford University, Stanford, CA 94305. E-mail: [email protected] 4 Professor, Computer Science Department, Stanford University, Stanford, CA 94305. E-mail: [email protected] 5 Group Leader, Manufacturing Systems Integration Division, National Institute of Standards and Technology, Gaithersburg, MD 20899. E-mail: [email protected]