| Summary: | 碩士 === 國立交通大學 === 資訊管理所 === 90 === Proxy server plays a very important role for internet access. It acts as an agent and resource manager for clients. Cache is key component of proxy server. It determinates the efficiency of proxy server. If we can take human behaviors into consideration, and integrate them into the design of a cache system, we may improve the performance of the cache system. In the paper, we propose an user-behavior-based prefetch mechanism to enhance the cache system. By predicting user behaviors, we may reduce the user’s idle time. This mechanism is based on the theory of habitual domain. Users are divided into four categories: “New User”, “Focus Accessing”, “Quick Glancing”, and “Internet New Type”. For the purpose of verifying the proposed mechanism, we develop a proxy server to evaluate the performance of the mechanism. By examing experimental results, we find that the user-behavior-based prefetch mechanism can reduce the average user waiting time by 36 percent, and increase the cache hit ratio of the cache system by 10 percent.
|
|---|
