2006 · 39 citations · 10 references
ComputingComputer ScienceDistributed SystemsProgramming LanguagesSchedulingDistributed ComputingScalable Algorithm DesignStream ProcessingWorkflow DesignWorkflow ManagementWorkflow OrchestrationData Stream ProcessingDistributed Stream ProcessingMetricsResource Consumption ModelingScalable Planning AlgorithmsStream Processing MetricsStream Processing Scalability
Recently the problem of automatic composition of workflows has been receiving increasing interest. Initial investigation has shown that designing a practical and scalable composi-tion algorithm for this problem is hard. A very general com-putational model of a workflow (e.g., BPEL) can be Turing-complete, which precludes fully automatic analysis of com-positions. However, in many applications, workflow model can be simplified. We consider a model known as the Stream Processing Planning Language (SPPL), applicable in stream processing and other related domains. SPPL replaces the no-tion of concurrency by timeless functional computation. In addition, SPPL defines workflow metrics of resource con-sumption and quality of service. Experiments have shown earlier that even a naı̈ve SPPL planning algorithm signifi-cantly outperforms existing metric PDDL planners on stream processing workflow composition problems. In this paper we describe an efficient and scalable algorithm for finding high-quality approximate solutions for large instances of SPPL problems. We demonstrate the scalability of the algorithm on synthetic benchmarks that are derived from practical prob-lems. We also give an example of SPPL model for practical problems.
10
Artificial intelligence and grids: workflow planning and beyond
Yolanda Gil, Ewa Deelman, James M. Blythe et al. · IEEE Intelligent Systems · 2004 · 144 citations
Prashant Doshi, Richard Goodwin, Rama Akkiraju et al. · International Journal of Web Services Research · 2005 · 112 citations