Edge and Fog Computing for Autonomous Network Management: A Comprehensive Survey

Fog computing is becoming popular as a way to reduce the bottlenecks that occur in computing and networking when deploying large numbers of IoT devices. Edge computing utilizes autonomous network management by performing local data processing (at the edge or source) which reduces both latency and network congestion. This also enables fog computing to provide a significant addition to cloud computing (where compute, network, store, and accelerate resources are located at the edge and network tiers) through the use of distributed, cooperative, and multi-tiered deployment of these elements. The purpose of this survey is to provide one of the first comprehensive reviews of edge and fog computing technologies for supporting autonomous network management. It discusses the role that Artificial Intelligence and Machine Learning play within the context of closed-loop control systems, intent-based networking, and zero-touch operations across the fog/edge layers of the network. The application areas considered in this report include 5G and future generations of mobile networks, the IoT, software-defined networking, and network function virtualization. For each of these application areas, specific challenges related to distributed resource management, scalability, interoperability, security, privacy, and reliability are also identified and discussed in detail.