IEEE Information Theory Chicago Chapter – DL Talk – Demand-private and Content-secure Caching Networks
September 18, 2020 @ 11:00 am - 12:00 pm CDT
Abstract: Caching is known to significantly reduce the communication load over bottleneck links by jointly designing the cache content of the users and the signal transmitted by the server so as to guarantee multicasting opportunities regardless of the requested content. In this talk we focus on two practically motivated constraints: (i) the content of the library must be kept secure from a wiretapper who obtains the signal sent by the server, and (ii) any subset of users together cannot obtain any information about the demands of the remaining users. We first revisit coded caching without privacy or security constraints [arxiv:1209.5807]. We then discuss how the linear function retrieval framework [arXiv:2001.03577] can be used as a building block in demand-private schemes as a means to protect the delivered content with locally cached privacy keys [arXiv:2008.03642]. Finally, we introduce the notion of key superposition, of privacy and security keys, to also guarantee content security [arXiv:2009.06000]. Remarkably, we show that the tradeoff between memory and communication load does not increase compared to the best-known schemes that only guarantee content security [arXiv: 1312.3961] or demand privacy [arXiv: 1908.10821].