Modeling and optimizing the cache deployment with filter effect in multi-cache system

Content-Centric Network is a new and promising architecture with in-network caching. In such a system with universal caching, the cache would play as a high-cut low-pass filter and have a big influence on the requests distribution, and further on cache efficiency. This paper looks into the change of request distribution though caches and presents a model of the “filter effect” in LRU (Least Recently Used) cache. Through the simulation in two common topologies, the experiment results have verified the validity of the new model. Since our model can quantify the network load of any certain cache deployment scheme, we optimize the way of deploying cache when the total cache size is limited. Simulation results show our optimal way can effectively minimize the total network load.