Proactive cache management method for content hash based distributed archive system

There are some distributed storage systems with linked data structure based on the content hash ID mechanism. In such systems, the user needs to wait to traverse multiple links to reach the requested data. Too slow total response time may degrade the usability of application. If the cache system is running on the user´s site, it can help to reduce the total response time of user´s request. But, there is no cache management mechanism that is applicable to such link type data structure. In this paper, we propose a novel management mechanism suitable for content hash ID based distributed storage system. The proposed mechanism is based on the expanded reference counter. We explain it as a concrete design for WOODS, which is a simple archiving data management system we have proposed. We evaluate the proposed cache mechanism using a measured data of Gnutella queries.