The Research of Automatic Data Hoarding Technique Based on Hyper Graph

The paper presents a low-cost automatic data hoarding technique. Data hoarding is the process of pre-fetching important data into local cache for future operation before disconnection. It can find the data sequence rules of broadcasting history according to data mining method, and formulate hype graph model and it can also cover the data of invalid window through low-cost hoarding computation. The theoretical analysis and experiment results indicate that the technique is low-cost and can solve cache pollution and can significantly improve cache hit rate and shorten average response time.