Optimize Block-Level Cloud Storage System with Load-Balance Strategy

Cloud storage systems take advantage of distributed storage technology and virtualization technology, to provide virtual machine clients with customizable storage service. They can be divided into two types: distributed file system and block level storage system. Orthrus is a Light weighted Block-Level Cloud Storage System, which adopt multiple volume servers´ architecture to avoid single-point problem in other solutions. However, how to make the servers load balance turn into a new problem appears in this architecture. In this paper we present a dynamic load balance strategy between multiple volume servers. We characterize machine capability and load quantity with black box modeling approach, and implement the load balance strategy based on genetic algorithm. Extensive experimental results show that the aggregated I/O throughputs of ORTHRUS are remarkably improved (about two times) with multiple volume servers, and both I/O throughputs and IOPS are remarkably improved (about 1.8 and 1.2 times respectively) by our dynamic load balance strategy.