The Swarm scalable storage system

摘要

Swarm is a storage system that provides scalable, reliable, andcost-effective data storage. Swarm is based on storage servers, ratherthan file servers; the storage servers are optimized forcost-performance and aggregated to provide high-performance data access.Swarm uses a striped log abstraction to store data on the storageservers. This abstraction simplifies storage allocation, improves fileaccess performance, balances server loads, provides fault-tolerancethrough computed redundancy, and simplifies crash recovery. We havedeveloped a Swarm prototype using a cluster of Linux-based personalcomputers as the storage servers and clients; the clients access theservers via the Swarm-based Sting file system. Our performancemeasurements show that a single Swarm client can write to two storageservers at 3.0 MB/s, while four clients can write to eight servers at16.0 MB/s