Metadata management in global distributed storage system

Data-intensive applications require the efficient management and transfer of terabytes or petabytes of information in wide-area. Data management system in distributed computing environment could not meet this requirement efficiently. Storage visualization is an efficient way to solve these problems. It integrates all kinds of high performance storage system to a unit one. It can not only share resources and make fully use of resources, but efficiently avoid conflict of data explosion and limited storage ability. To resolve these conflicts, we propose a Global Distributed Storage System (GDSS). With data explosion, metadata becomes more and more enormous. Exinting methods could not keep up with the need of scalability, availability and efficiency to manage metadata. To solve these problems, MDC (MetaData Controller), a novel metadata management scheme based on MatchTable, is introduced in GDSS. In MDC, MatchTable is responsible for communication between SSP (Storage Service Point) and DS (Directory Server), and cache module is responsible for increasing of metadata access efficiency and as an assistant to maintain the replica coherence of metadata. MDC locates metadata only once by matching MatchTable, which avoids searching it from root directory server step by step. Experiment results indicate that MDC can locate metadata fast and supply fault tolerance, and also has better availability and scalability.