Potential students projects for an elective course or CS397/CS497

Note: Some BTech projects or Master's level projects may be toned down for a class project or a CS397/CS497. Please talk to me about this.


Project title: BER vs. SNR measurement for commercial 802.11 receivers
Project description: This project involves the measurement of Bit-Error-Rate (BER) or Packet-Error-Rate (PER) at various Signal-to-Noise-Ratio (SNR) levels for a given 802.11 receiver (PCMCIA card, or an Access Point).
Background required: Basic networking (CS), fundamentals of communication (EE). These can also be learnt along the way.
Project level: Class project for a senior-level course, elective, or CS497
Project variations: Different receivers may be considered. Each can be a two-person project.
Project title: TCP Performance on top of an STDMA MAC
Project description: This project involves the study of TCP performance on top of an STDMA-based MAC. See here for a description of STDMA and the context.
Background required: Computer networking; use of ns-2 simulator can be learnt along the way.
Project level: Class project or CS497
Project variations: There are many variations of this project possible, and each can be considered as a separate project. (1) You can study different network configurations, each with a single TCP connection, (2) You can study multiple TCP connections. There are probably 2-3 class projects possible in this topic. Please talk to me for more details.

Project title: Performance of TCP over multiple (wireless) interfaces: simulation
Project description: With different wireless technologies becoming popular, a mobile client may have many wireless interfaces which can be used simultaneously. However, simultaneous use of multiple communication paths presents problems in the presence of TCP. This is due to packet reordering across the multiple communication paths. One approach to address this issue is to use Mobile-IP, and run multiple TCP connections, one per wireless interface between the Home-Agent and the Mobile. Thus a TCP connection between a server and the mobile gets split into three different TCP connections at the Home-Agent. The purpose of this project is to study the performance on TCP under such conditions, using a simulation.
Background required: Computer networking; knowledge of ns-2 simulator is desirable, but can be learnt along the way.
Project level: Class project or CS497
Suggested by: Kameswari Chebrolu

Project title: Performance of TCP over multiple (wireless) interfaces: implementation
Project description: The description is the same as the corresponding simulation project described above, except that you will be implementing the setup, instead of using the ns-2 simulator.
Background required: Computer networking; socket programming is desirable, but can be learnt along the way.
Project level: Class project or CS497
Suggested by: Kameswari Chebrolu

Project title: Study of router power consumption
Project description: Router power consumption has not received much attention so far, although it is quite important in the following context. In rural networking, there is a lot of value to reducing router power consumption so that routers can be powered using solar energy. This project involves studying this issue of router power consumption. This would involve study of some literature on this, and may be some measurements as well. This can be figured out along the way. Your study may be immediately applicable in the Digital Gangetic Plains project.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of network emulation platforms
Project description: There are three methodologies for network system/protocol evaluation: simulation, emulation, and implementation. The advantage of a simulation is that it gives control over the experiment(s) and repeatability. Simulation also allows one to explore a wide range of possibilities quickly. On the other hand, simulations are sometimes not realistic, and do not capture all system parameters, such as processing and memory bottlenecks. Also, if one is going to implement the system/protocol anyway, a simulated implementation is duplicated effort. On the other end of the spectrum is an implementation-based evaluation. While this may be realistic, often times, the experiments may not be repeatable since all parts of the (distributed) system may not be under the control of the experimenter. A network-emulation platform combines the advantages of both. It is better than a simulation in that it captures all parameters of the system. It is better than an implementation in that it gives better control over the experiment. This project would involve surveying past/current research on network emulation platforms.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of BGP security issues
Project description: BGP (Border Gateway Protocol) is the inter-domain routing protocol on the Internet today. When BGP was originally designed, there were no security mechanisms built-in. Hence there is no way to check if a routing announcement is genuine or fake. Malicious (or misconfigured) BGP route announcements can redirect traffic, or "black-hole" traffic, disrupting large parts of the Internet. This is currently a hot area of research. This project involves surveying current research on this topic.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of research on Denial-of-Service (DOS) attacks on the Internet
Project description: Denial-of-Service (DOS) attacks have become very common on the Internet today. There has been a lot of research on detecting such attacks, and some counter-measures as well. This project involves surveying research on this topic..
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of research on BGP stability, convergence, and performance
Project description: The Border Gateway Protocol (BGP) is the inter-domain routing protocol on the Internet today. Being a path-vector protocol, it has poor convergence properties. That is, it can take a long time to recover after a network failure. Apart from the theoretical poor convergence, there are also several poor convergence properties due to poor implementation. There are been a lot of interest on studying BGP convergence properties, and performance under scale. This project would involve surveying the recent research on this topic.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of research on Quality-of-Service (QoS) in 802.11 networks
Project description: The 802.11 contention-based MAC is not suited for achieving network Quality-of-Service (QoS). That is, it is difficult to give bandwidth or delay guarantees for audio/video applications that require such guarantees. There has been recent research on how such support can be provided. This project would involve surveying the current efforts to this end.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of research on power-saving routing in sensor networks
Project description: Sensor networks are networks formed by various sensors, for instance, for collection of data on a field. Since the sensors operate without any external source of power, saving energy is a very important issue. This project involves surveying research on this topic.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of research on secure routing
Project description: Another aspect of sensor networks (continuing from above) is that the routing needs to be secure since the sensor nodes are out on the open field. This project will involve a survey of techniques for secure routing in this context.
Background required: Computer networking.
Project level: Class project or CS497

Project title: Survey of routing, MAC, or transport on adhoc networks
Project description: There has been plenty of research on ADHOC routing, and some recent research on MAC protocols and transport protocols in ADHOC networks. This project would involve surveying one of these three areas of research -- three different student projects can consider the three areas.
Background required: Computer networking.
Project level: Class project or CS497

Bhaskaran Raman
Last modified: Mon Aug 4 00:22:36 IST 2003