Abstract: Data centers are the vital infrastructure platform for big data and cloud computing applications, and meanwhile the efficiency of the data center network will determine the processing capacity of the data center. The dynamic and burstiness characteristics of services in time and space domains cause congestions in data center networks, resulting in low network resources utilization and poor quality of service for users. The path criticality based congestion avoidance rerouting method is proposed in this paper, where the link criticality metric related to link load and delay performance is designed to represent the influence of large flows on the link. Then, the characteristics of large and small flows in data center networks are perceived, the rerouting path allocation problem is modeled with the objective of minimizing the maximal link criticality, and a heuristic rerouting algorithm based on path criticality is proposed to alleviate the current congestions. The simulation results have demonstrated that the proposed method can improve the packet loss rate, throughput and delay performances for data center networks.