Drizzle: Cooperative Symbol-Level Network Coding in Multichannel Wireless Networks

Abstract

Errors are inherently present in unreliable wireless channels. The primary challenge in designing error-control protocols in the medium-access control (MAC) or physical layer is to effectively maximize achievable throughput in wireless networks, even when unpredictable and time-varying errors exist. Network coding has successfully been applied to improve throughput in IEEE 802.11-based wireless networks with a shared broadcast channel. In state-of-the-art physical-layer designs in multichannel wireless networks [such as IEEE 802.16 Worldwide Interoperability for Microwave Access (WiMAX)], however, the convenience of a shared wireless broadcast channel to perform opportunistic listening no longer exists, and hybrid automatic repeat request (HARQ) is the predominant error-control protocol in the physical layer, rather than plain automatic repeat request in IEEE 802.11 MAC. Would network coding be well employed in multichannel wireless networks and able to bring further improvements over HARQ? This paper proposes Drizzle, which is a new solution to maximizing throughput with the presence of errors, which takes advantage of network coding at the symbol level in multichannel wireless networks. By operating at the symbol level and using soft-decision values, we show that Drizzle is able to exploit both time and cooperative diversity in realistic multichannel wireless networks, to adapt to time-varying and bursty channel errors, and to efficiently collect as many correct symbols as possible at the receiver.