Collaborative Service Caching for Edge Computing in Dense Small Cell Networks

TL;DR

This paper introduces a decentralized collaborative service caching algorithm for dense small cell networks in mobile edge computing, optimizing service placement to improve system performance while addressing heterogeneity and demand coupling.

Contribution

It proposes a novel decentralized algorithm based on parallel Gibbs sampling for collaborative service caching in MEC, with extensions for selfish base stations using coalitional game theory.

Findings

The CSC algorithm significantly improves caching efficiency and convergence speed.

The approach guarantees optimality and stability in collaborative caching.

Extensions incentivize cooperation among selfish base stations.

Abstract

Mobile Edge Computing (MEC) pushes computing functionalities away from the centralized cloud to the proximity of data sources, thereby reducing service provision latency and saving backhaul network bandwidth. Although computation offloading has been extensively studied in the literature, service caching is an equally, if not more, important design topic of MEC, yet receives much less attention. Service caching refers to caching application services and their related data (libraries/databases) in the edge server, e.g. MEC-enabled Base Station (BS), enabling corresponding computation tasks to be executed. Since only a small number of services can be cached in resource-limited edge server at the same time, which services to cache has to be judiciously decided to maximize the system performance. In this paper, we investigate collaborative service caching in MEC-enabled dense small cell (SC) networks. We propose an efficient decentralized algorithm, called CSC (Collaborative Service Caching), where a network of small cell BSs optimize service caching collaboratively to address a number of key challenges in MEC systems, including service heterogeneity, spatial demand coupling, and decentralized coordination. Our algorithm is developed based on parallel Gibbs sampling by exploiting the special structure of the considered problem using graphing coloring. The algorithm significantly improves the time efficiency compared to conventional Gibbs sampling, yet guarantees provable convergence and optimality. CSC is further extended to the SC network with selfish BSs, where a coalitional game is formulated to incentivize collaboration. A coalition formation algorithm is developed by employing the merge-and-split rules and ensures the stability of the SC coalitions.