Paper ID | N.3.3 | ||
Paper Title | Extending the Optimality Range of Multi-Antenna Coded Caching with Shared Caches | ||
Authors | Emanuele Parrinello, Petros Elia, EURECOM, France; Eleftherios Lampiris, Technical University of Berlin, Germany | ||
Session | N.3: Coded Caching I | ||
Presentation | Lecture | ||
Track | Networking and Network Coding | ||
Manuscript | Click here to download the manuscript | ||
Virtual Presentation | Click here to watch in the Virtual Symposium | ||
Abstract | This work considers the cache-aided multiple-input single-output broadcast channel (MISO BC) where an $L$-antenna transmitter serves $K$ receiving users, each assisted by one of $\Lambda < K$ caches with normalized capacity $\gamma$. For this setting it was known that in the range of $L \in [K/\Lambda,K\gamma]$, one can achieve a Degrees-of-Freedom (DoF) performance of $L+K\gamma$. This very restrictive constraint that $L\leq K\gamma$ excluded shared-cache settings with a large number of antennas from achieving the maximum DoF; for the more realistic regime of $L>K\gamma$, all existing coded caching schemes suffer substantially reduced caching or multiplexing gains. Our work provides a novel coded caching scheme that achieves the exact best known, near optimal, DoF $L+K\gamma$, and does so even if $L>K\gamma$, thus covering an important hole in identifying the optimal performance for the multi-antenna shared-cache problem. Therefore, our work reveals that shared-cache systems with many transmit antennas can also enjoy both full multiplexing gains ($L$) as well as full caching gains $(K\gamma)$ despite the sharing of the caches. A side benefit of this scheme is its applicability in multi-antenna settings with dedicated users caches, where it can offer the advantage of reducing the subpacketization without sacrificing the DoF performance. |
Plan Ahead
2021 IEEE International Symposium on Information Theory
11-16 July 2021 | Melbourne, Victoria, Australia