WLBS: A Weight-Based Metadata Server Cluster Load Balancing Strategy
Metadata Server is a critical component of object-based storage system, which is used to manage file system name space, direct user access and map files and directories to physical storage devices. Existing metadata server cluster load balancing strategies fail to consider the case of heterogeneous metadata server cluster, causing a potential performance bottleneck. In this paper, the authors introduce WLBS, a dynamic load balancing strategy for heterogeneous metadata server cluster. For each of the metadata server in the cluster, WLBS can precisely evaluate its inherent metadata processing capability by a performance model based on IO latency.