In-network caching is an appealing solution to cope with the increasingbandwidth demand of video, audio and data transfer over the Internet.Nonetheless, an increasing share of content delivery services adopt encryptionthrough HTTPS, which is not compatible with traditional ISP-managed approacheslike transparent and proxy caching. This raises the need for solutionsinvolving both Internet Service Providers (ISP) and Content Providers (CP): bydesign, the solution should preserve business-critical CP information (e.g.,content popularity, user preferences) on the one hand, while allowing for adeeper integration of caches in the ISP architecture (e.g., in 5G femto-cells)on the other hand. In this paper we address this issue by considering a content-obliviousISP-operated cache. The ISP allocates the cache storage to various contentproviders so as to maximize the bandwidth savings provided by the cache: themain novelty lies in the fact that, to protect business-critical information,ISPs only need to measure the aggregated miss rates of the individual CPs anddo not need to be aware of the objects that are requested, as in classiccaching. We propose a cache allocation algorithm based on a perturbedstochastic subgradient method, and prove that the algorithm converges close tothe allocation that maximizes the overall cache hit rate. We use extensivesimulations to validate the algorithm and to assess its convergence rate understationary and non-stationary content popularity. Our results (i) testify thefeasibility of content-oblivious caches and (ii) show that the proposedalgorithm can achieve within 10\% from the global optimum in our evaluation.
展开▼