Assignment handed out on Sunday, March 11, 2001

Final Project Overview

Your ultimate goal is to write and present a proposal that wins an award from the aD Uni funding board. You will work in the same pairs that have already been assigned. The process will proceed as follows: The proposal + the presentation counts for 20% of your total grade. Your grade will depend on how well you convey your ideas and how well you answer questions from your peers and the TA.

After the presentations are delivered on 3/21, each recitation section will be responsible for choosing 1 team to represent them at the finals on 3/22. This will be done via a vote. Each member of the recitation section (including your TA) will have 1 vote. In the case of a tie, a second vote will determine who goes on. In the second vote, those team members involved in the tie will not vote. If the vote is still tied, the TA's vote will be used to select the winner.

The final presentations will be on 3/22. These presentations are delivered by the teams voted to represent their respective recitation sections. The presentation will be 40 minutes in length, with 30 minutes of questions from the aD Uni funding board (the staff, TA's, and a few others). These presentations will account for the remaining 20% of the final grade for each member of the respective recitation sections' grades. As before, the grade will depend on how well your representatives convey their ideas and how well questions from the final judgement board and the aD Uni students are answered.

Your Proposal

Your proposal will be a response to an NSF RFP around Network Middleware. Your proposal must describe your approach to developing a sofware layer that addresses the conflicting issues of: For this project, you must choose an application you wish to optimize. A rule of thumb: choose an application where the impact of fully optimizing the application significantly impacts network performance. The less that optimizing your application impacts network performance, the easier your proposed solution will be to develop. The tradeoff is that your proposal will be easier to shoot down. The more that optimizing your application impacts network performance, the more difficult your proposal solution will be to develop. In this case, a good proposed solution will make your proposal easier to defend.

Take telnet sessions, for example. As we discussed in lecture, building session support at the lowest network layers can cause huge performance issues for streaming protocols. TCP is a "middleware" layer that supports sessions with some tradeoffs. TCP does not affect streaming protocols because streaming protocols can rest on UDP.

Contact the teaching staff with any questions.