Professors Jeff Kennington and Eli Olinick have received $163,550 in funding from the Office of Naval Research to study Using Linear Programming for Optimal IP Routing. A description of the project is as follows.
The current design of routers and router software inhibits the use of a central processor and mathematical programming software to optimize the operation of a data network. Currently routers use the network topology and link weights to determine routing tables using a distributed algorithm. The routing strategy must then adhere to what is known as the equal-cost multi-path split rule. We propose modifying the router software by allowing them to accept routing tables and relieving them of the burden of the equal-cost multi-path split rule. We believe that these simple modifications will permit traffic engineers to optimize the operation of large data networks. That is, linear programming can then be used to help manage these sophisticated systems. We will develop new algorithms to improve the existing routing problem as well as one based on our new proposal.