Machine Learning (ML) has been enjoying an unprecedented surge in applications that solve problems and enable automation in diverse domains. Primarily, this is due to the explosion in the availability of data, significant improvements in ML techniques, and advancement in computing capabilities. Undoubtedly, ML has been applied to various mundane and complex problems arising in network operation and management. There are various surveys on ML for specific areas in networking or for specific network technologies. This survey is original, since it jointly presents the application of diverse ML techniques in various key areas of networking across different network technologies. In this way, readers will benefit from a comprehensive discussion on the different learning paradigms and ML techniques applied to fundamental problems in networking, including traffic prediction, routing and classification, congestion control, resource and fault management, QoS and QoE management, and network security. Furthermore, this survey delineates the limitations, give insights, research challenges and future opportunities to advance ML in networking. Therefore, this is a timely contribution of the implications of ML for networking, that is pushing the barriers of autonomic network operation and management.

Over the last decade, a significant amount of effort has been invested on architecting agile and adaptive management solutions in support of autonomic, self-managing networks. Autonomic networking calls for automated decisions for management actions. This can be realized through a set of pre-defined network management policies engineered from human expert knowledge. However, engineering sufficiently accurate knowledge considering the high complexity of today's networking environment is a difficult task. This has been a particularly limiting factor in the practical deployment of autonomic systems. ML is a powerful technique for extracting knowledge from data. However, there has been little evidence of its application in realizing practical management solutions for autonomic networks. Recent advances in network softwarization and programmability through SDN and NFV, the proliferation of new sources of data, and the availability of lowcost and seemingly infinite storage and compute resource from the cloud are paving the way for the adoption of ML to realize cognitive network management in support of autonomic networking. This article is intended to stimulate thought and foster discussion on how to defeat the bottlenecks that are limiting the wide deployment of autonomic systems, and the role that ML can play in this regard.

Traditional routing protocols employ limited information to make routing decisions, which can lead to a slow adaptation to traffic variability, as well as restricted support to the Quality of Service (QoS) requirements of applications. This paper introduces a novel approach for routing in Software-Defined Networking (SDN), called Reinforcement Learning and Software-Defined Networking Intelligent Routing (RSIR). RSIR adds a Knowledge Plane to SDN and defines a routing algorithm based on Reinforcement Learning (RL) that takes into account link-state information to make routing decisions. This algorithm capitalizes on the interaction with the environment, the intelligence provided by RL and the global view and control of the network furnished by SDN, to compute and install, in advance, optimal routes in the forwarding devices. RSIR was extensively evaluated by emulation using real traffic matrices. Results show RSIR outperforms the Dijkstra’s algorithm in relation to the stretch, link throughput, packet loss, and delay when available bandwidth, delay, and loss are considered individually or jointly for the computation of optimal paths. The results demonstrate that RSIR is an attractive solution for intelligent routing in SDN.

Network Slicing and Deep Reinforcement Learning (DRL) are vital enablers for achieving 5G and 6G networks. A 5G/6G network can comprise various network slices from unique or multiple tenants. Network providers need to perform intelligent and efficient resource management to offer slices that meet the quality of service and quality of experience requirements of 5G/6G use cases. Resource management is far from being a straightforward task. This task demands complex and dynamic mechanisms to control admission and allocate, schedule, and orchestrate resources. Intelligent and effective resource management needs to predict the services’ demand coming from tenants (each tenant with multiple network slice requests) and achieve autonomous behavior of slices. This paper identifies the relevant phases for resource management in network slicing and analyzes approaches using reinforcement learning (RL) and DRL algorithms for realizing each phase autonomously. We analyze the approaches according to the optimization objective, the network focus (core, radio access, edge, and end-to-end network), the space of states, the space of actions, the algorithms, the structure of deep neural networks, the exploration–exploitation method, and the use cases (or vertical applications). We also provide research directions related to RL/DRL-based network slice resource management.

Traditional routing protocols employ limited information to make routing decisions, which leads to slow adaptation to traffic variability and restricted support to the quality of service requirements of applications. To address these shortcomings, in previous work, we proposed RSIR, a routing solution based on Reinforcement Learning (RL) in Software-Defined Networking (SDN). However, RL-based solutions usually suffer an increase in time during the learning process when dealing with large action and state spaces. This paper introduces a different routing approach, called Deep Reinforcement Learning and Software-Defined Networking Intelligent Routing (DRSIR). DRSIR defines a routing algorithm based on Deep RL (DRL) in SDN that overcomes the limitations of RL-based solutions. DRSIR considers path-state metrics to produce proactive, efficient, and intelligent routing that adapts to dynamic traffic changes. DRSIR was evaluated by emulation using real and synthetic traffic matrices. The results show that this solution outperforms the routing algorithms based on Dijkstra’s algorithm and RSIR in relation to stretch, packet loss, and delay. Moreover, the results obtained demonstrate that DRSIR provides a practical and feasible solution for routing in SDN.

Network slicing is a promising technology for providing customized logical and virtualized networks for the fifth-generation (5G) use-cases (enhanced mobile broadband, ultrareliable low-latency communications, and massive machine-type communications), which pose distinct quality of service (QoS) requirements. Admission control and resource allocation mechanisms are pivotal for realizing network slicing efficiently, but existing mechanisms focus on slicing the radio access network. This article proposes an approach encompassing intelligent and efficient mechanisms for admission control and resource allocation for network slicing in the 5G core network. The admission control mechanism introduces two solutions, one based on reinforcement learning (called SARA) and the other based on deep reinforcement learning (called DSARA). SARA and DSARA consider the QoS requirements of 5G use-cases, differentiate network core nodes from edge nodes, and process slice requests in time windows to favor the service provider’s profit and resource utilization. Results show that SARA and DSARA overcome existing mechanisms for managing admission control and resource allocation in 5G core network slicing.