Stochastic Dynamic Cache Partitioning for Encrypted Content Delivery

In-network caching is an appealing solution tocope with the increasing bandwidth demand of video, audioand data transfer over the Internet. Nonetheless, an increasingshare of content delivery services adopt encryption throughHTTPS, which is not compatible with traditional ISP-managedapproaches like transparent and proxy caching. This raises theneed for solutions involving both Internet Service Providers(ISP) and Content Providers (CP): by design, the solutionshould preserve business-critical CP information (e.g., contentpopularity, user preferences) on the one hand, while allowing fora deeper integration of caches in the ISP architecture (e.g., in5G femto-cells) on the other hand.In this paper we address this issue by considering a content-oblivious ISP-operated cache. The ISP allocates the cache storageto various content providers so as to maximize the bandwidthsavings provided by the cache: the main novelty lies in the factthat, to protect business-critical information, ISPs only need tomeasure the aggregated miss rates of the individual CPs anddoes not need to be aware of the objects that are requested, asin classic caching. We propose a cache allocation algorithm basedon a perturbed stochastic subgradient method, and prove thatthe algorithm converges close to the allocation that maximizes theoverall cache hit rate. We use extensive simulations to validatethe algorithm and to assess its convergence rate under stationaryand non-stationary content popularity. Our results (i) testifythe feasibility of content-oblivious caches and (ii) show that theproposed algorithm can achieve within 10% from the globaloptimum in our evaluation