Networking Over White Spaces (KNOWS)

Established: December 19, 2008

The next generation of wireless networks will include software defined radios, cognitive radios, and multi-radio systems which will co-exist harmoniously while operating over a very wide range of frequencies. Under the umbrella of the KNOWS project we are revisiting “classical” wireless networking problems and designing new solutions that incorporate and build upon recent advances in software and hardware technologies for networking over the recently opened white spaces spectrum.

Brief Description

The WhiteFiService APIs and web front-end can be accessed by clicking here.
Check out deployment pictures by clicking here.

The white spaces spectrum is fundamentally different from the ISM bands where Wi-Fi operates along three main axes. First, it exhibits spatial variation since a channel available at one node might be occupied by a primary user (TV, microphone) at another node in the network. Second, the spectrum is not contiguous. Some channels might be occupied by primary users hence causing the spectrum to be fragmented. Finally, there is temporal variation since an available spectrum might be occupied at a later time by a primary user, e.g. wireless microphone.

Given these challenges, we have researched several techniques to form networks over this part of the spectrum. In the first version of KNOWS we introduced the concept of Time Spectrum Blocks (TSB) as the fundamental unit over which two nodes could communicate.  We designed a control-channel based medium access control protocol, called CMAC, for enabling nodes with different spectrum views to access the medium. Associated with CMAC we proposed and evaluated an algorithm, called bSMART, for efficiently allocating the spectrum to different contending nodes.

In the second version of KNOWS, we looked at the problem of setting up a base station in the white spaces spectrum. In this system, called WhiteFi, we eliminated the need for a dedicated control channel. In addition, we proposed a new technique, called SIFT, that enables nodes to rapidly discover base stations operating at different center frequencies using different channel widths by analyzing signals in the time domain. We proposed and evaluated a new metric, called MCham, using which the base stations choose the “best” chunk of the spectrum to operate on, where the spectrum chunk can span multiple channels. We have prototyped this system on Windows.

People

Portrait of Ranveer Chandra

Ranveer Chandra

Managing Director, Research for Industry

Portrait of Thomas Moscibroda

Thomas Moscibroda

Distinguished Engineer Azure Core Platform Capacity & Efficiency Microsoft Azure

Portrait of Victor Bahl

Victor Bahl

Technical Fellow & Chief Technology Officer, Azure for Operators

Portrait of Tusher Chakraborty

Tusher Chakraborty

Senior Researcher in M365 Copilot