Abstract : —In this paper, we study cache-enabled small cell networks (SCNs) with local regularly requested content sampling to take into account local user interests for the cache decisions. We consider Zipf-like local content popularity with variables indicating the correlation level of user interests in the same region. Based on stochastic spatial models for the small cell base station (SCBS) and user distribution, we provide analytical results on the cache service probability, i.e. the probability that an arbitrary user finds its requested content cached in its nearby SCBSs. The tradeoff between the service probability and the sampling cost is discussed and the optimal sampling range given by maximizing the service probability under the cost constraint is derived. Numerical results with different correlation levels of local user interests are given, which validate our analysis on the service-cost tradeoff.
