Coded caching can exploit multicast opportunities even when multiple users request different pieces of content, and thus can significantly reduce the backhaul requirement to serve high-volume content. A common assumption in existing studies of coded caching is that all files are with the same size, which however may not be true in reality. Our previous work [1] first studied this problem, and proposed a non-trivial lower bound, as well as a new achievable scheme that uses a caching probability increasing proportionally with the file size. However, the gap [1] of the achievable rate and the lower bound still differs by a factor of Θ(log K), where K is the number of users in the system. In this paper, under a mild assumption that the total size of all files is larger than eight times the size of one individual cache, we will close this gap and reduce it to a constant by proposing a novel new lower bound and another new achievable scheme. Our lower bound is derived by considering a new cut-set bound, where files of a type1 will be requested more often if the number of such files is smaller. Our achievable scheme uses a caching probability that decreases with the number of files with a same type. The improvements on both the lower bound and the achievable rate make their gap constant.
[1]
Suhas N. Diggavi,et al.
Hierarchical coded caching
,
2014,
2014 IEEE International Symposium on Information Theory.
[2]
Urs Niesen,et al.
Decentralized coded caching attains order-optimal memory-rate tradeoff
,
2013,
2013 51st Annual Allerton Conference on Communication, Control, and Computing (Allerton).
[3]
Jaime Llorca,et al.
Order-Optimal Rate of Caching and Coded Multicasting With Random Demands
,
2015,
IEEE Transactions on Information Theory.
[4]
R. Pulikkoonattu.
ON THE INEQUALITIES IN INFORMATION THEORY
,
2007
.
[5]
Urs Niesen,et al.
Coded Caching With Nonuniform Demands
,
2017,
IEEE Transactions on Information Theory.
[6]
Xinbing Wang,et al.
Coded caching for files with distinct file sizes
,
2015,
2015 IEEE International Symposium on Information Theory (ISIT).
[7]
Urs Niesen,et al.
Fundamental Limits of Caching
,
2014,
IEEE Trans. Inf. Theory.
[8]
Suhas N. Diggavi,et al.
Coded Caching for Multi-level Popularity and Access
,
2014,
IEEE Transactions on Information Theory.