A coalitional game approach for fractional cooperative caching in content-oriented networks

摘要

In content-oriented networks, popular contents are replicated at the intermediate nodes to enhance content delivery performance. Under cooperative caching, the caching nodes collaborate to leverage one another's cache capability and to reduce the amount of traffic transferring inside the network. This study considers the cooperation among service providers (SPs). The transferable-payoff coalitional game model is applied for analysis. We investigate the stability of the grand coalition and show that the dual-based cost allocation is in the core. A linear program (LP) minimizing the network bandwidth-expense is used for the characteristic function of the game model. However, solving the LP is a challenge because of a large amount of contents in the network. The Dantzig-Wolfe decomposition approach is further applied to decompose the large-scale problem into many subproblems, which can be solved in parallel. The analysis provides not only a deeper insight into the cooperative cache among SPs but also content placement and distribution strategies as a solution to the LP.