On the network sharing of mixed network coding and routing data flows in congestion networks

In this paper, we study the congestion game for a network where mutliple network coding (NC) and routing users sharing a single common congestion link to transmit their information. The data flows using NC and routing will compete network resources, and we need to determine the optimal allocation of network resources between NC and routing data flows to maximize the network payoff. To facilitate the design, we formulate this process using a cost-sharing game model. A novel average-cost-sharing (ACS) pricing mechanism is developed to maximize the overall network payoff. We analyze the performance of ACS in terms of price of anarchy (PoA). We formulate an analytical expression to compute PoA under the ACS mechanism. In contrast to the previous affine marginal cost (AMC) mechanism, where the overall network payoff decreases when NC is applied, the proposed ACS mechanism can considerably improve the overall network payoff by optimizing the number and the spectral resource allocation of NC and routing data flows sharing the network link.