Motivated by sensor network deployments on bridges and in habitats, we have been studying the problem of rate control in multi-hop wireless networks. Our focus has been to explore practical distributed mechanisms for fair and efficient rate control. In this talk, I will begin by describing IFRC, a rate control technique suitable for the constrained communication patterns found in sensor networks.
I will then describe more recent work on rate control in a more generalized setting, that of a multi-hop mesh network. In this work, motivated by the well-known problem of TCP flow starvation in multi-hop wireless networks, we ask what mechanisms would be necessary to ensure a fair and efficient congestion control in mesh networks. Specifically, we explore two classes of congestion control: AIMD-based designs, and those based on providing explicit feedback to sources.