An Efficient Adaptive Fault Diagnosis Algorithm for Highly Scalable Data Center Networks

Xiangke Wang, Yudu You, Xiao-Yan Li, Ximeng Liu, Yang Yang

Abstract: The big data system based on data center network provides low-latency, high-quality services for big data applications. When server failure occurs in data center network, the security of the big data platform and the service quality of big data applications will be severely affected. A highly scalable data center network (HSDC) is an emerging server-centric data center network that achieves incremental scalability while ensuring low cost and energy consumption, low diameter, and high bisection width. In this paper, we determined the connectivity and diagnosability of HSDC. Then we firstly design an efficient adaptive fault diagnosis algorithm to diagnose the actual status of all servers in HSDC with at most \({m2^m+4m(m-2)}\) (resp. 9) tests, where m is the dimension of the HSDC and \(m \ge 3\) (resp. \(m = 2\)). Experimental results show that for HSDC, our algorithm can achieve complete diagnosis and greatly reduce the number of tests.