Development of an interactive auction system in a local area network environment
An interactive Simulated Auction Application System (SAAS) has been designed and implemented for the Department of Economics to study auction behavior analysis for market decision making. SAAS can be used to conduct six types of auctions, each with four variations. The application is built on a local area network of twelve IBM personal computers interconnected by the IBM Cluster hardware and software. A new KERMIT file transfer protocol with sliding window has also been designed and implemented to transfer data collected from the SAAS to a mainframe for statistical analysis. The performance of the new KERMIT is nearly three times better than the original one with respect to response time and channel utilization. Valuable experience is gained in designing communication protocols for personal computers and in studying their performance.