Fully heterogeneous demand-private coded caching beyond two users and two files

Characterize the memory–rate trade-off for information-theoretic demand-private coded caching with heterogeneous file sizes, heterogeneous cache capacities, and arbitrary user-demand distributions for general numbers of files and users beyond the two-file, two-user setting.

Background

The paper studies demand-private coded caching in a fully heterogeneous setting, where file sizes and cache capacities may differ and user demands may follow arbitrary non-uniform distributions. It establishes an exact memory–rate characterization only for the fundamental case of two files and two users, using an extension of the virtual-user and restricted-demand-subset techniques.

The authors explicitly identify the broader fully heterogeneous demand-private coded caching problem as unresolved for general system parameters. Solving this problem would extend the paper’s two-user, two-file characterization to settings with more files and users and would determine the impact of heterogeneity and demand privacy in general.