Potential Master's projects


Project title: Simultaneous reception in 802.11 using directional antennae
Project description: The setup for this project is as follows. There are two 802.11 point-to-point links setup using directional antennae.  There are thus four end-points.  Two of the end-points are physically close together.  In other words, they are at the same "node".  The aim of this project is to study the feasibility of simultaneous reception at this common node, along the two different directions.
Background required: Basic networking (CS).  Knowledge of 802.11 and antennae desirable, but can be picked up along the way.
Project level: Master's or Bachelor's project depending on depth of exploration

Project title: Realistic support for 802.11 in ns-2
Project description: Without going too much into the details (which you can talk to me about in person), ns-2 is a simulator built for simulating various kinds of networks.  It is extensible in that one can add code to it to make it simulate things it was not originally built for. The simulator does not have very good support for directional antennae, channel interference modeling and so on. Significant code has been built to add support for this by Ankur Khandelwal and Siddartha Saha (BTech graduates from IITK), although this is not complete. This project would involve providing full support for 802.11. Another aspect of the project would be to show that the modeling done is in fact correct, by matching up simulation results with collected traces.
Background required: Basic networking, wireless networking, C++ programming background, TCL background helps, but can be learnt along the way. Familiarity with ns-2 is desirable, but can be learnt along the way.
Project level: Bachelor's or Master's project depending on scope
Project status: Ashwini and Sabyasachi have taken this up for their BTP.
Project title: Edge colouring in a graph, with angular constraints
Project description: Edge colouring in a graph is the problem of assigning colours to edges.  A proper edge colouring is one in which no two adjacent edges receive the same colour.  In this project, we will deal with a slightly different constraint.  We will consider graphs which are derived as follows.  A set of points are chosen on a two-dimensional plane, and edges are straight-lines between the nodes.  Note that this need not necessarily be a planar graph in that edges may cross one another.  Thus at a node, any two edges have an angular separation between them in the range 0-180 degrees.  Our constraint for edge colouring is that no two adjacent edges within an angular separation of alpha degrees should receive the same colour.  The goal is to characterize this problem to the extent possible.  You may make assumptions about only nearby nodes in the two dimensional plane being neighbours, or about the graph being planar, or the maximum degree in the graph being a small number (4-5).  You can also make reasonable assumptions about alpha.  This problem has applications in channel allocation for point-to-point 802.11 wireless links.  Talk to me in person for more details as to what assumptions are practically feasible.
Background required: Graph theory. Networking background helps in understanding the problem setting, but is not really necessary.
Project level: Master's or Bachelor's project depending on scope

Project title: Construction of planar-point graph, with angular constraints
Project description: In the same setting as the above problem, suppose we have the freedom to construct the graph edges.  That is, we are given a set of nodes as points in a two-dimensional plane. We are to choose edges such that (a) the graph has "reasonable" connectivity (e.g. graph is connected in the least), and (b) no two adjacent edges are within alpha degrees of each other.  Again, this problem has applications in construction of topologies with point-to-point 802.11 wireless links. Talk to me in person for further details.
Background required: Graph theory. Networking background helps in understanding the problem setting, but is not really necessary.
Project level: Master's or Bachelor's project depending on scope

Project title: Secure routing in a wireless mesh network out in the open fields
Project description: In a wired network, the routers are physically secure, and hence protected against tampering. However, in a wireless ISP network with nodes out in the fields, there are two problems. Anyone can snoop on, or even send spurious packets. What is more, the routers are not physically secure. This project involves analyzing the security of a routing protocol in this context. That is, come up with possible attacks in a naive system, and solutions against those attacks, with may be reasonable assumptions.
Background required: Computer and network security, Computer networking.
Project level: Master's or Bachelor's project depending on scope
Project status: Paul Ipe has taken this up for his BTP.

Bhaskaran Raman
Last modified: Sat Aug 2 15:35:43 IST 2003