Proposal: CNS 0551692 PI: Sabharwal, Ashutosh Institution: Rice University
The principal investigators at Rice University will develop the Wireless Open-Access Research Platform (WARP) that promotes a holistic and rapid approach to wireless network design. WARP will be a scalable and extensible platform with three component layers: custom hardware with scalable processing and extensible I/O, platform support packages that provide seamless integration across different hardware components, and an application design environment. They will develop an open-access repository with WWW access that allows WARP users to construct wireless networks, share experiments over the Internet, and implement their networks on their own WARP hardware kits; lastly the investigators will develop a wireless development kit and make it available to other researchers and educators, conduct workshops on use of the WARP system, and host a student exchange program.
The main goals for our Wireless Open-Access Research Platform (WARP) were to design a highly flexible, capable and extensible wireless platform for clean design of network protocols. Our driving goal was allow ease of access for academic researchers to such a point that they could perform any wireless experiment with effortlessness and in a scientific manner. Further, the teams' objective was to ensure that the hardware was universally available and all software designs were fully open-source. We exceeded every objective laid out in the original proposal. Most salient outcomes have been: Outreach Activities: Over 100 research organizations worldwide have adopted WARP for their wireless research and prototyping projects. WARP users now include researchers in major industrial labs, domestic and international universities and universities abroad. As of November 2011, the distribution of organizations with WARP is as follows: Number of US Universities = 29 Number of International Universities = 45 Number of US Corporate Labs = 15 Number of International Corporate Labs = 15 Total = 104 As clear from the above data, WARP impact is truly global and has been instrumental in establishing completely new labs, which would not have ventured into experimental research otherwise. Workshops: The Rice WARP team continues to teach free workshops for wireless researchers. Eleven workshops have been held so far, at Rice University, Nile University in Cairo, Egypt and IIT Madras in Chennai, India, University of Oulu, Finland, National Chiao-Tung University in Taiwan and RWTH in Germany. All workshops are open to wireless researchers in both academia and industry, and all of the workshop materials are made available open-source on the WARP site. Cumulatively, more than 300 students, faculty and corporate researchers have been trained in these workshops. Most workshop participants are now active WARP users. Community Development: The community of WARP users continues to grow as more groups adopt the platform. The primary means of community interaction are public forums hosted by the Rice WARP project, with more than 4300+ posts. The forums are used primarily by current WARP users seeking support from the Rice WARP team and by potential WARP users seeking more information about the platform. More recently, the forums have been used by both Rice and outside users to discuss possible extensions to existing WARP designs and other wireless research project ideas. This expansion from one way tech support to interactive research collaboration reflects the increasing sophistication of early WARP adopters. Their in-house expertise is enabling ever more complex projects with many sharing their expertise via the forums for the benefit of the overall community. WARP not only made high-accuracy clean wireless research possible but enabled some significant breakthroughs in wireless systems. Most interestingly, the wide availability of the hardware and open-source nature of the platform ensured there were multiple research groups engaged in these new breakthrough areas. The two significant breakthroughs came in the form of challenging the two basic assumptions in all wireless networks. First, all wireless networks assume that the nodes can only send or receive in the same band at any given time, i.e, the nodes are half-duplex in a given frequency band. We showed, using WARP, that this assumption is not always valid and in fact, single-channel, full-duplex wireless communications can be enabled reliably for many systems. Second, all mobile nodes, like smartphones, are designed to use omni-directional antennas. That is because, till date, directional transmissions on small form-factor devices have been considered impossible. However, using WARP, it was recently shown that a power-efficient directional communication on small form-factor is possible and can achieve large gains in actual networks.
