The overall goal of the project was to analyze the performance of client/server programming using a wide variety of network configurations. Performance was measured by response time and traffic generation on the network. Response time was defined as the average time span between submission of a request and receipt of a response. Traffic generation was concerned with the number of packets generated and an analysis of the traffic pattern on both single LAN segments and on multiple LAN segments.

The student researchers met 9-12 hours weekly. Each week the students and faculty met to discuss project milestones and problems encountered during the past week. We began by ensuring that our physical machine setup was adequate to meet our needs and was working properly. Our next step was to write and revise the applications needed to test the network and to analyze the log files crated by the testing applications. The last step was to collect and analyze the data.

We collected data on a variety of configurations including:

1. a single client and server running on the same machine;


2. a single client and server running on separate machines;


3. one server and multiple clients, each on separate machines connected over a single LAN;


4. one server and multiple clients distributed over two LANs connected by a router;


5. multiple clients and two servers running on the same LAN;


6. multiple clients and two servers running on two LANs connected by a router with each server on a different LAN segment.

The six configurations were divided into three experiments. Configurations 1 and 2 used a client/server chat program. Configurations 3 and 4 used a datagram monitoring system. Configurations 5 and 6 used a set of programs designed to simulate a three-tiered server setup. The results obtained were very interesting and sometimes non-intuitive.

Although frustrating at times, our efforts to determine why actual results did not always match expected results provided an excellent learning experience. One of the major benefits of this experience was to learn how important careful planning and a well thought out research methodology are in tackling a project of this magnitude.

As a final step in the project, we were able to share our research results when we were invited to make a poster presentation for The International ACM Student Research Contest, held at SIGCSE 2000 Technical Symposium. We also presented our results at a local student research forum, Celebration of Quality, on the Northwest Campus.

We thank CREW for their support of this project.