Fault-Tolerant Video Analysis Cloud Scheduling Mechanism

Abstract

This paper proposes a fault tolerant optimal scheduling mechanism based on dependency analysis to make large-scale video surveillance systems more efficient and more reliable. We study large-scale video surveillance systems and associated theoretical issues, and then analyze the efficiency and reliability that application obtained in large-scale video surveillance systems under constrained circumstance through decision analysis and optimization theory, and finally validate and optimize the scheduling model through experimental analysis. The experiments indicate the proposed scheduling mechanism can effectively improve the global analysis job completion ratio and ensure the efficiency and reliability of the large-scale video surveillance network.