Our portion of this work has been supported by the National Science Foundation under Grants No. ECS-0300130 and CCF-0430892.
As ubiquitous wireless networks become a reality, it is imperative for physical layer researchers to consider end-to-end network performance while prescribing the physical layer, and for network layer researchers to exploit contemporary physical layer capabilities to optimize network performance. We have worked with UMass networking researcher Don Towsley in this area. In [Liu/Goeckel/Towsley, GlobeCom 2002], we set the rate of an adaptive physical layer to optimize end-to-end performance of a TCP session running over the link. More recently, we have considered: (1) how physical layer advances, particularly cooperation among nodes, can change network level behavior [Song/Goeckel/Towsley, InfoCom 2006], (2) how network layer algorithms can exploit contemporary physical layer degrees of freedom [Vasudevan/Zhang/Goeckel/Towsley, InfoCom 2006], and (3) limits of the improvement in scaling performance provided by network coding in large ad hoc networks [Liu/Goeckel/Towsley, InfoCom 2007]. The latter work is of particular interest, demonstrating that network coding with a standard point-to-point physical layer will not be able to significantly improve the pessimistic scaling results of Gupta/Kumar.