DRSIR: A Deep Reinforcement Learning Approach for Routing in Software-Defined Networking

Abstract

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 the 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 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 the Dijkstra's algorithm and RSIR, in relation to stretching (stretch), packet loss, and delay. Moreover, the results obtained demonstrate that DRSIR provides a practical and viable solution for routing in SDN.

Cognitive routing in software defined networks using learning models with latency and throughput constraints

“…During the exploitation phase of the RNN, the selection of the next node in the NFE process is influenced by the individual excitation levels of each neuron. To determine neuron i's potential (qi), we can use (1).…”

“…In the task of managing and running computer networks, the introduction of software defined networks (SDNs) marks a significant shift in strategy. SDN is a design approach that isolates the control plane (which decides on routing) from the data plane (which sends packets) [1]. Because of this split, network managers may more easily monitor and control network resources in a centralised, programmable fashion, which has several advantages.…”

A new intelligent cross-domain routing method in SDN based on a proposed multiagent reinforcement learning algorithm

“…However, this method faces challenges related to constructing triples, ensuring the best overall migration efficiency under the current triple and shortening the time overhead of iteration. Considering that traditional routing method uses some information to implement forwarding operations, which cannot achieve rapid adaptation to traffic variability and ensure the QoS of the network, Casas-Velasco et al. (2021) proposed a DRL and SDN intelligent routing (DRSIR) method that considers state metrics such as the available bandwidth, latency and loss path priority to generate efficient intelligent routing paths adapted to variable traffic and used natural and synthetic traffic matrices to evaluate DRSIR through a simulation, significantly reducing the loss of packets and lowering the network delay cost.…”

A New Intelligent Cross-Domain Routing Method in SDN Based on a Proposed Multiagent Reinforcement Learning Algorithm

“…However, this method faces challenges related to constructing triples, ensuring the best overall migration efficiency under the current triple, and shortening the time overhead of iteration. Considering that traditional routing uses some information to implement forwarding operations, which can not achieve rapid adaptation to traffic variability and ensure the QoS of the network, Casas-Velasco et al 33 proposed a DRL and SDN intelligent routing (DRSIR) method that considers state metrics such as available bandwidth, latency and loss path priority as the best routing to generate efficient intelligent routing adapted to variable traffic, and using natural and synthetic traffic matrix to evaluate DRSIR through simulation, significantly reduce the loss of packet and reduce the network delay cost. However, this method does not consider the influence of the future trend in network traffic state on network performance.…”