Cache-aided Function Retrieval with Security and Privacy Constraints
Sponsors:
Problem Statement and Motivation
- Caching is an effective way to smooth out network traffic in peak times by storing locally content during peak-off times.
- Caching has two phases: placement (limited only by the size of the local cache and designed without knowledge of future demands) and delivery (where the goal is to minimize the number of broadcast transmissions so that every receiver can decode the desired message)
- In this work we study how users can retrieve functions from possibly distributed servers while maintaining user- and server-privacy as well as security from eavesdroppers.
Technical Approach
- We extend tools for the classical shared single bottleneck-link caching problem to networks with legitimate users (interested in retrieving a function of the files at the servers) and eavesdroppers.
- We derive achievable and converse bounds that are to within a constant gap of one another.
Key Achievements and Future Goals
- We derived state-of-the-art achievable and converse bounds, which meet for some network parameters.
- Results presented at:
- Qifa Yan and Daniela Tuninetti, Key Superposition Simultaneously Achieves Security and Privacy in Cache-Aided Linear Function Retrieval, arXiv:2009.06000, ITW2020 to appear
- Qifa Yan and Daniela Tuninetti, Fundamental Limits of Caching for Demand Privacy against Colluding Users, arXiv:2008.03642, and IEEE Journal on Selected Areas in Information Theory
- Kai Wan, Hua Sun, Mingyue Ji, Daniela Tuninetti, Giuseppe Caire, Fundamental Limits of Device-to-Device Private Caching with Trusted Server, arXiv:1912.09985, and ICC 2020, and ISIT 2020
