REARED PARSIMONIOUS FILE SYSTEM BY CLOUD

With simple approach interfaces and flexible billing models, cloud storage has become an interesting solution to simplify the storage management for both business and individual users. However, old file systems with large advantage for local disk-based storage backend cannot fully accomplish the internal features of the cloud to obtain good performance. In this paper, we present the design, implementation, and decision of Coral, a cloud based file system that beat a balance between performance and financial cost. Unlike previous studies that treat cloud storage as just a normal backend of existing networked file systems, Coral is designed to address several key issues in improving cloud-based file systems such as the data layout, block management, and billing model. With carefully designed data structures and algorithms, such as analyzing semantically correspond data blocks, KD(k- Dimentional tree based caching policy with self-adaptive thrashing prevention, useful data layout, and excellent garbage collection, Coral achieves good performance and cost savings under various workloads as demonstrated by large evaluations.