arrow
Return

Generalized hierarchical coded caching

delete2024-12-01
delete0
PRE
AI
M
Manuel Fernández‐Veiga
F
Francisco Troncoso‐Pastoriza
DOI:10.1016/j.jnca.2024.104027delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
Optimizing data traffic is a key concern in content distribution networks to reduce the bandwidth needed to serve the requested content to the final users. In this context, hierarchical coded caching has been proposed as an effective method for reducing traffic. The typical two-level scenario consists in a tree-like structure: on the first layer, a set of intermediate nodes or helpers with local caches store and serve content to a subset of users at the second layer, who also have their local caches. A central server partially stores its content between the first- and second-layer caches, in uncoded form, and subsequently transmits coded messages to satisfy the files demanded by the users. This topology has been thoroughly analyzed in the literature over the past years; however, little attention has been put in arbitrary topologies, where users might be connected to multiple helper nodes, which is especially interesting for scenarios such as wireless networks. Thus, in this paper we study a hierarchical two-level coded caching system with an arbitrary topology and propose a novel content distribution method for this generalized model that leverages the multiple connections between users and helpers to improve the performance of the distribution process. We give achievability results for this system and compare its performance with other classical single-layer and two-layer coded caching systems. Our results show that further gains are possible by using two levels of caching in different operating regimes.
Keywords:
Coded caching
Hierarchical networks
Shared chaches
Multicast networks

Journal

Journal of Network and Computer Applications cover
Journal of Network and Computer Applications
IF:
8
Papers:
3.6K
Citations:
1.1W

Organization

A
atlanttic
Scholars:
101
Papers: 84
Citations: 0
U
Universidade de Vigo
Scholars:
7.7K
Papers: 8.3K
Citations: 13