High reliability and flexibility storage cluster in decentralized organization

Decentralized, cooperative and large-scale distributed storage systems that consist of a cluster of storage nodes attached with local disks can deliver high resource utilization, high availability and easy scalability. Therefore they become a cost-effective solution. High availability with node failures, strong data consistency, flexible expandability and high I/O throughput are still important challenges in the face of the system design. In this paper, we describe the design and implementation of a LAN-based decentralized storage cluster built upon off-the-shelf commodity components. It employs replication and erasure-codes to provide high reliability and automatic failure recovery, adapts versioning to achieve single-file serializability and replicas consistency, supports data migration scheme to balance I/O workload and storage utilization, and provides global unified content cache to optimize memory-to-disk policy. We also present the experiment results to demonstrate the features and performance of our design