When Serverless Computing Meets Edge Computing: Architecture, Challenges, and Open Issues

Edge computing enables applications to leverage computing resources near the data source to perform data processing, which can effectively cope with the explosive growth of network data. However, along with the benefits come great challenges: application developers need to explicitly manage their resources and handle the burden of scalability and load balancing; users have to pay based on the allocated resources, not the resources actually consumed; and so forth. Serverless edge computing extends the idea of serverless computing and has emerged as a compelling model for dealing with many of the challenges associated with edge infrastructure. Different from the existing works which mainly focus on the basic architectures, platforms, and optimization models, this article aims to provide a systematic and comprehensive overview of server-less edge computing networks from the perspective of networking. We first propose the network architecture and layered structure of serverless edge computing networks. Then, the communication process, as well as the implementation and deployment, is presented. Next, the promising technical challenges, including service deployment and lifecycle management, resource awareness and service discovery, service scheduling, and so on, are discussed. Finally, some potential areas for future research are highlighted.

Introduction

With the rapid development of new mobile applications, such as the Internet of Things (IoT), augmented reality (AR), and autonomous driving, mobile data generated from various distributed sources shows an explosive growth trend [1]. To cope with such an amount of data, cloud computing with beneficial properties such as on-demand self-service, resource pooling, and rapid elasticity, has become a promising approach [2]. However, since the cloud servers are generally located far away from users, the latency introduced by sending/retrieving heavy payloads from/to the cloud can be prohibitive. Edge computing has the potential to alleviate these costs and performance bottlenecks [3]. This computing paradigm enables applications to leverage computing resources in close proximity to data sources to perform data processing before forwarding the results to other nodes and cloud servers, reducing bandwidth usage and speeding up application response.