基于逆向分层的网格工作流调度算法.pdfVIP

31 2 Vol. 31 No. 2 2008 2 CHINESE JOURNA L OF COMPU TERS Feb. 2008 1) , 3) 1) , 2) 1) , 2) 1) , 2) 苑迎春 李小平 王 茜 张 毅 1) ( 2100 6) 2) ( 2100 6) 3) ( 071001) DAG( Directed Acrylic Graph) . DAG , ( Bottom Level, BL) , , DBL( Deadline Bottom Level) . DTL( Deadline Top Level) , , , . DBL M CP ( minimum Crit ical Path) DTL , DTL MCP 1562% , DBL MCP 2474% . . ; ; ; ; TP3 3 Bottom Level Based Heuristic for Workflow Scheduling in Grids YUA N YingChun1) , 3) LI XiaoPing1) , 2) WANG Qian1) , 2) ZHANG Yi1) , 2) 1) (School of Comp uter Science and Engineer ing , Southeast University , N anj ing 2100 6) 2) (Key Laboratory of Computer N et ork and Inf ormation I ntegration (Southeast University) , Ministry of Education, N anj ing 2100 6) 3) (Faculty of I nf ormation Science and Technology , A griculture University of H ebei, Baoding, H ebei 071001) Abstract Efficient scheduling of w orkflow applications represented by DAG ( Directed Acyclic Graph) w ith the objective of timecost optimization is fundamental and intractable in computa tional Grid. By analyzing parallel and synchronization properties of DAG, all tasks are divided in to BL ( Bottom Level) groups using a backw ard method. T he w orkflow deadline can be trans formed into the time intervals of all tasks based on BL groups, a bottom leveling based algorithm called DBL ( Deadline Bottom Level) is proposed. In DBL, the starting time of a