Due to the growing number of World Wide Web users, improving the Quality of Service (QoS) of the Web, especially reducing user access latency, is important to both Internet Service Provides and users. User access latency is a serious problem for international and low bandwidth connection users.
Some technologies, including web caching, replication, pushing, and prefetching have been explored by many research groups. Most of these studies focus on proxy servers and web servers, but little work has been done to improve the performance of clients.
This work combines user specification, network geography information, web content type, communication delay, and other issues to create a system which can be used to generate an optimized prefetch schedule for a particular user. We can predict the user's access schedule and prefetch the related web objects to the local cache to reduce the user web access latency.
Testing shows an obvious QoS improvement using prefetching. Cache hit rates have increased 34.8% to 49.6%. Response times have been reduced 7.16% for high bandwidth connection users and 30.1% for low bandwidth connection users. Prefetching has proved helpful in reducing access latency and improving the QoS, especially for clients with low bandwidth connections or high access regularity.