On Balancing Client-Server Load in Intelligent Web-Based Applications Involving Dialog

With the explosive growth of the Internet have come problems of increased server load and network latency. This means that systems that require interchange of data between server and client over the network can be slow to unusable (at peak times). We are investigating methods for developing distributed CBR systems which minimise the load on the servers and the network, thus increasing response time and usability. The first part of this research focuses on balancing load in a client/server situation that involves a long-lived dialog. This report discusses the various issues to be tackled when attempting to effectively balance the processing load. Introduction This report summarises the work to date in one strand of the CBR-Net project. This project is in the area of networked/distributed Case-Based Reasoning. Case-Based Reasoning (CBR) is a methodology found in Artificial Intelligence, which takes solutions to problems encountered in the past as a basis for solving current problems. The..