A Migration Approach to Optimal Control of Wireless Networks

(NSF Grant CNS-1524317)

PI: Eytan Modiano

Recent growth in mobile and media-rich applications has greatly increased the demand for wireless capacity, straining wireless networks. This dramatic increase in demand poses a challenge for current wireless networks, and calls for new algorithms that make better use of scarce wireless resources.  Recently developed algorithms for optimally managing wireless resources hold the promise of significant performance improvement, but require all of the nodes in the network to be upgraded with new functionality – which is both costly and impractical.  This project introduces a novel architectural paradigm for wireless networks, whereby optimal algorithms are designed to operate in networks with both new and legacy nodes.  This new paradigm allows optimal algorithms to be incrementally deployed alongside existing schemes, thus providing a migration path for new control algorithms, and the promise of dramatic improvement in network performance at incremental cost.

This project develops a novel overlay architecture for implementing optimal network control algorithms over legacy networks. New nodes, capable of implementing sophisticated network control algorithms, will be connected in a virtual network overlay that operates on top of the legacy network.  The research will answer fundamental questions about which nodes must be upgraded with new functionality and the tradeoff between the number of new upgraded nodes and network performance.  The project will develop new routing algorithms that send packets from their sources to their destinations in the overlay network, and transmission scheduling algorithms for mitigating the effect of wireless interference. These new algorithms will be designed to operate efficiently in a network with a mix of new and legacy nodes, by taking interoperability into account.  Thus, this project will answer fundamental questions about the introduction of new control techniques into legacy networks, and provide a promising approach to bridging the gap between new techniques developed for universal deployment and the reality of the networks in operation today.

Our research goals include:

  • development of an overlay architecture for implementing optimal network control algorithms over a legacy network, using a limited number of overlay nodes.
  • development ofnew routing and flow control algorithms that operate over the overlay network.
  • development of optimal scheduling algorithms for a wireless network with a mix of controllable and uncontrollable nodes.
  • validation of our algorithms using a combination of simulation, emulation, and testbed implementation. 

Prof. Eytan Modiano, PI
Thomas Stahlbuhk, graduate student

Related Publications:

  1. Nathan Jones, George Paschos, Brooke Shrader, Eytan Modiano, “An overlay architecture for Throughput Optimal Multipath Routing,”  in IEEE MobiHoc 2015, extended version submitted to ToN.
  2. Georgios  Paschos and Eytan Modiano, “Throughput optimal routing in overlay networks,” Allerton conference on Communication, Control, and Computing, September 2014.