Cooperative Network Coding and Coding-Aware Channel Assignment in Multi-Channel, Multi-Interface Wireless Networks

Efforts to improve the capacity of multi-channel, multi-interface wireless mesh networks have mainly focused on mitigating channel interference and balancing traffic loads. In a limited number of channels and interfaces, this approach cannot help encountering the network overload and traffic saturation, as network traffic increases. It will be serious, especially at intersecting nodes such as the nodes around gateways for last-mile connectivity. Considering this situation, which degrades the aggregate throughput of networks, a more aggressive strategy to cope with network traffic saturation is necessary. We propose a novel cooperative network coding scheme, which exploits coded-overhearing, for unicast in multi-channel, multi-interface wireless mesh networks. Further, we present a coding-aware channel assignment algorithm with new metrics to support our network coding scheme, resulting in substantial improvement in the aggregate throughput. The combination of the proposed network coding scheme and the channel assignment algorithm contributes to overcoming geographical limitations in conventional network coding. It also shows better flexibility for the insufficient number of interfaces and the outage of coding opportunities. Our evaluation results show maximally a 52% improvement in terms of the aggregate throughput by using our coded-overhearing algorithm with a coding-aware channel assignment.