When Edge Caching Meets a Budget: Near Optimal Service Delivery in Multi-Tiered Edge Clouds

Abstract

More and more artificial intelligence (AI) applications, such as virtual reality (VR) and video analytics, is rapidly progressing towards enterprise and end users with the promise of bringing immersive experience. Such AI applications need unprecedented requirements for ultra-low latency. Edge computing is emerged to guarantee the timeliness for the delay-sensitive applications. The delay experienced by AI users can be significantly reduced, by caching various services that are initially deployed at data centers to cloudlets in edge networks. However, it is impractical to cache all services to cloudlets, due to often limited operational budget of service providers and resource capacity constraints of edge cloudlets. In this paper, we investigate a fundamental problem of service caching from remote data centers to cloudlets in a multi-tiered edge cloud network. We first develop two approximation algorithms with approximation ratios to solve the problem for a single type of service demanded by users. We then devise an efficient heuristic to solve the problem that users request different types of services. We finally conduct extensive experiments on a real testbed to evaluate the performance of the proposed algorithms, and experimental results demonstrate that our algorithms can outperform some existing algorithms significantly with 25% lower average delay.