Grid scheduling method based on hierarchical scheduling strategy and Dynamic Data replication

Grid scheduling method based on hierarchical scheduling strategy and Dynamic Data replication

Rai Jinhui Liangsong

Aiming at how to efficiently perform task scheduling and data replication in grid, so as to reduce task execution time, a task scheduling algorithm (YPP) and an optimized Dynamic Data replication algorithm (E1/ZF) are proposed, and a scheme is constructed to combine the two algorithms effectively. The scheme uses YPP algorithm to comprehensively consider the number of task waiting queue, the location of task demand data and the calculation capacity of the site, adopt the method of hierarchical scheduling of network structure, calculate the integrated task cost with appropriate weight coefficient, search the best computing node region, and analyze data transmission time and storage access delay with E1/ZF algorithm.　, waiting for the copy request in the storage queue and the distance between the nodes, selecting the best replica location in a large number of replicas, combined with replica placement and replica management, thus reducing file access time. The simulation results show that the proposed scheme shows better performance than other algorithms in terms of average task execution time.

Grid scheduling method based on hierarchical scheduling strategy and Dynamic Data replication