CS 552 Fall 2008

Class Project Ideas


The purpose of the class project is to introduce the research process to students. A successful project will propose and and evaluate an idea and/or discover and quantify an unknown property of an existing system. An example class project is here. For your initial 1 page write up describe: (1) Project team members, (2) The goal or end result of the project (3) How you will reach those goals, and (4) what unknowns are there in your plan to reach the project goals.


Students will submit a proposal, a midpoint report, and final project report. The 1 pages project proposals are due Friday, Sept. 26, at 5PM.


Below are possible project ideas.


  1. Adding Location Information to the network. These projects are investigate various aspects of adding position and location into the network infrastructure. For infrastructure for these projects, we have the GRAIL Real Time Location System running inside the 2nd and 3rd floors of the Core buildings. GRAIL can track Wifi Devices as well as Active RFID tags. In addition, you may use the OpenMoko phones as a platform for development.

    1. Rutgers Campus Navigator. The goal of this project is to develop a set of layers and interfaces to allow a hypothetical small device to navigate between any two rooms on campus. You should assume some indoor positioning system at the source and target buildings. A way to start this project would be to develop a prototype using the openmoko phones and the GRAIL infrastructure and then argue how such a prototype could be generalized.

    2. Rendezvous. Another application would be to enable rapid-ad hoc meetings of people. More formally, given N persons and their P positions, one could create an application that suggested a place to meet at one of L locations. As with the above project, you should create a small prototype and then argue how you would generalize the idea.

    3. Passive Mobility Detection. Recent work has shown that it is possible to use the wireless infrastructure to not only track wireless devices, but also to detect objects distorting the EM field generated by normal wireless traffic. In this project, you would try to replicate the results of this paper. You may also develop a more refined algorithm. We would provide you with some sniffers from the GRAIL system to use in this work.

    4. Enhancements to 802.11 to support localization. Better localization results than are possible using signal strength (e.g. RSSI modalities) could be obtained by exposing additional information in the physical layer. These projects explore how well end localization performance

      1. Fast clocks. The current 802.11 standard specifies a mico-second level resolution clock timestamp as part of every packet. Investigate how well the localization performance could be improved if this clock were increased in frequency to 10 Mhz, 100Mhz, or even a Ghz.

      2. Reliable echo. Take an existing 802.11 design and design a stable echo/ping. That is, the turnaround time should be constant. Such an echo/ping could used for ranging by subtracting out the fixed part and

  2. Network Tomography. These projects concern how to measure and develop models of the Internet via measurement and modeling.

    1. Using a sandwich probe experiment to deduce the internal structure of the Rutgers network. Compare the results to the actual network to evaluate the performance of the tomography. To build the structure of the network, you could use a Maximium Likelihood Estimate scheme discussed in class, or you could use a hierarchical clustering scheme discussed in this paper.

    2. Survey the Rutgers Networks using the techniques described in this paper. Compare the results to the those described for the USC network.

  3. Bandwidth Variability. Wireless PHY layers do a reasonable job of minimizing packet loss at a cost of decreasing the available bandwidth. However, this means that wireless clients will face a new problem of variable bandwidth in spite of the wireless' network ability to mitigate losses (Suggested by Tao Ye from Sprint Labs).

    1. Compare TCP vs. UDP comparison under fast/slow varying channel conditions, with a PF scheduler (to simulate a 3G wireless Questions to be answered could be -- is QoS necessary with a typical mixed workload? What can servers do for such mobile traffic? There could also be discussions on what metrics should be used to measure overall performance.

    2. Quantify the efficiency loss of TCP over links with variable bandwidth. You will need to abstract a changing link bandwidth (e.g. discreet jumps vs. continuous), define optimal performance, and then define a metric that captures the percentage difference of the realized bandwidth from the optimal. For this project you will also need to run some simple experiments that measure real TCP loads over controlled changing link bandwidths. .

  4. Cloud Computing. New demands will be placed on the network as more applications are built using 'cloud computing' models.

    1. Machine-level vs. Language-level virtualization. One method of providing infrastructure for cloud computing is by giving each user a full virtual machine. Another is by uploading language level code (e.g. java, PHP or ruby) and giving each user access to a database. Quantify the difference in resource demands for each approach. That is, givena web application, how much CPU, memory and bandwidth does it take using each approach? Also, describe the reliability and manageability differences between the two approaches.

    2. TBD.

  5. Others

    1. Quantification of the Discrepancy of Classic Queuing Theory. Many works have shown that network packet traffic follows self-similar behavior in contrast to exponential behavior. However, classic queuing networks are still the simplest, easiest to use models of computer networks available. The goal of this project is to quantify the error of simple queuing models when traffic follows self-similar behavior. A successful project would quantify the discrepancy of a variety of properties obtainable from classic queuing theory such as the average time spent in a queue, average queue length, and drop probabilities, as a function of the self-similarity of the traffic. This project would likely use a combination of theoretical analysis and trace-driven simulation.