Congestion price for cache management in information-centric networking

We tightly integrate caching and congestion control into a unified approach to enhance user-centric performance in information-centric networking. To do that, we take advantage of the available congestion price fed back by congestion control to guide caching decision at each content router. The point is to let caching work in a longer time scale to alleviate congestion retrieving expensive content in terms of congestion cost. As a consequence, not only does caching increase throughput for users who request the cached content, but also it contributes to a large extent towards reducing network congestion, which benefits all other users who share the same congested parts of the network. Our use of existing congestion feedback signals eliminates the need for out-of-band congestion measurement such as bandwidth share computation which in reality might not reflect the true congestion state of the delivery paths. The proposed method yields positive performance gain in simulations compared with existing cache schemes.