A data prefetching method applied to cloud workflow scheduling

Abstract

A data prefetching method applied to cloud workflow scheduling. In the workflow scheduling process, there is a sequence in the completion time of the execution of multiple precursor activities of the aggregation activity. When a precursor activity is executed, it can be transferred on the resource In the case of idle bandwidth, the output file after the execution of a certain predecessor activity is transferred to the resource where the common successor of the workflow scheduling process is located in advance. This optimization method is called data prefetching; in the workflow scheduling process, steps S1 and S2 Both are data prefetching, the difference is that the timing of data prefetching is different; data prefetching is performed at the resource allocation moment of the predecessor activity; if an activity is the predecessor activity of a certain aggregation activity, then after allocating resources to the predecessor activity, The target resources need to be set, and the output files or data of other completed predecessor activities are transferred to the set target resources in advance.

Description

technical field [0001] The invention belongs to the technical field of cloud workflow, in particular to a scheduling optimization technology in the technical field of cloud workflow. The invention can effectively improve resource utilization and workflow execution efficiency. Background technique [0002] In the workflow management system, the scheduling module is one of its most important core components. Its main goal is to assign workflow activities to corresponding resources through appropriate scheduling strategies based on cloud computing resource information and workflow information. . However, due to the heterogeneity and dynamics of resources, the workflow scheduling problem becomes extremely complex, and the quality of the scheduling algorithm directly affects the execution time of activities and the utilization rate of resources. Therefore, how to design an efficient scheduling method is one of the important research directions of workflow. Since the workflow sc...

AI Technical Summary

Problems solved by technology

[0005] Among the above algorithms, the Min-Min, Max-Min and Sufferage algorithms can be classified as activity-based priority scheduling algorithms, because when calculating the activity priority, the algorithm only considers the parameters of the activity itself, and does not consider the entire workflow , while the HEFT, CPOP and DCP-G algorithms consider the impact of critical paths on workflow scheduling from a global perspective. Although HEFT and CPOP algorithms use critical path technology, their critical paths are determined before scheduling. The weight of the activity is also fixed

Images