Fault tolerant storage and data access optimization in data center networks

Abstract

Data center, of which performance affects the cloud computing, is the key part of cloud computing technology. As cloud computing has penetrated into our daily life, the data stored in data center have increased rapidly, which increases the requirements of fault tolerance of distributed storage system. In the meantime, the efficiency of data access should be considered when applying network coding scheme in data center networks. This paper tackles these two issues. The first part of this paper proposes a fault-tolerant storage based placement strategy. We consider three factors: the efficiency of data access, the load of storage node, and the expectations of recovery time of failure nodes. When the loads of nodes are balance and recovery time for failure nodes is shorter, we can achieve a high fault-tolerance storage in data center networks. Simulation results show that the fault-tolerant placement has low expectation of recovery time and can achieve load balance. In the second part of this paper, we design a priority-queue based scheduling policy of data access to improve the efficiency of data access when deploying network coding in data center networks. Simulation results show that the proposed priority queue scheduling policy can achieve low delay by reducing the number of frozen servers.