SEARS: Space Efficient And Reliable Storage System in the Cloud

Commercial cloud storage services today must be designed to handle large amount of data with the dual aim of fast data retrieval and storage reliability without sacrificing storage cost. We propose SEARS, a cloud-based storage system which integrates data deduplication and erasure coding schemes in a flexible fashion to support reliable and efficient data storage with fast user response time. By properly associating data with storage server clusters, SEARS offers various server binding schemes and the flexibility to mix different configurations in the system for various application scenarios, making it suitable for both archival and real-time data storage needs. Evaluation of our prototype implementation of SEARS over Amazon EC2 shows that it outperforms existing storage systems in storage efficiency, data reliability and file retrieval time. SEARS delivers retrieval time of 2.5 s for 3 MB files compared to 7 s in existing systems.