USA - 2017
For creative contributions to wireless networking.
Over the past 20 years, wireless networks have become the predominant mode of network communication for people and their devices. The world is now awash in wireless signals. Dina Katabi has produced a string of exceptional results by applying methods from communication theory, signal processing, and machine learning to solve problems in wireless networking, and to use wireless signals to develop new ways to sense human activity.
Katabi and her collaborators have written several papers on overcoming interference in wireless networks, which occurs when multiple nodes transmit concurrently. These papers model how signals mix in the air as a code, and describe algorithms that allow the receiver to decode such interference-based codes. These methods not only overcome interference, but may even be exploited to increase network throughput.
Over the past few years, Katabi and her students pioneered the use of wireless signals in sensing applications, particularly the ability to "see through" walls with Wi-Fi signals. They showed how to sense humans, track their movements, and measure their vital signs using the reflection of radio signals off their bodies. This work expands the concept of radar and introduces new algorithms to deal with the complexity of indoor multi-path radio reflections, the deformative nature of the human body, and the weak and unknown interaction between vital signs such as human heartbeats and radio signals.
USA - 2013
For contributions in cross-layer wireless networking, wireless network coding, and Internet congestion control.
USA - 2012
For her seminal contributions to the theory and practice of network congestion control and bandwidth allocation.
Dina Katabi's dissertation work on the explicit Control Protocol (XCP) is one of the most important contributions in the area of network congestion control. It has introduced theoretical methods in particular, stability analysis from control theory to the design of scalable practical network protocols. Such protocols aim to minimize network congestion, thereby maximizing utilization efficiency, while ensuring a fair allocation of capacity among different flows that compete for the same bandwidth. Katabi's XCP was the first protocol to achieve both goals simultaneously without imposing the impractical requirement that routers maintain per-flow state. Her innovation is based on the novel and theoretically justified observation that congestion control mechanisms and capacity allocation among flows can be addressed separately. She designed an ingenious packet-marking scheme to carry information about network congestion and used methods from control theory to show that her algorithm provided a stable solution for suitable values of the algorithm's parameters to the delayed-feedback control problem defined by this scheme. Her work thus solved an important practical problem in a novel and rigorous way while opening up hitherto unexplored connections between networking and control theory.
Katabi's research on XCP initiated a new approach to network protocol design and is essential reading for current courses in the area. Her ideas have impacted many subsequent protocols and computer systems. Perhaps more importantly, by changing the way we think about the algorithmic control of network behavior, Katabi's work is likely to impact future Internet developments to an even greater extent.
USA - 2003
For her dissertation, Decoupling Congestion Control from the Bandwidth Allocation Policy and its Application to High Bandwidth-Delay Product Networks, nominated by the Massachusetts Institute of Technology.