CoDRL: Intelligent Packet Routing in SDN Using Convolutional Deep Reinforcement Learning
| Title | CoDRL: Intelligent Packet Routing in SDN Using Convolutional Deep Reinforcement Learning |
| Publication Type | Conference Paper |
| Year of Publication | 2019 |
| Authors | Swain, P., Kamalia, U., Bhandarkar, R., Modi, T. |
| Conference Name | 2019 IEEE International Conference on Advanced Networks and Telecommunications Systems (ANTS) |
| Date Published | Dec. 2019 |
| Publisher | IEEE |
| ISBN Number | 978-1-7281-3715-5 |
| Keywords | CoDRL model, Convolution layer, Convolutional Deep Reinforcement Learning, convolutional layer, convolutional neural nets, coupled congestion control, current SDN systems, Deep Deterministic Policy Gradient (DDPG), Deep Deterministic Policy Gradients deep agent, deep reinforcement learning, deep reinforcement learning agent, dynamic packet routing, dynamic traffic engineering, flexible traffic engineering, gradient methods, intelligent packet routing, less efficient resource utilization, mean network delay, network congestion, network data, network performance, Packet loss rate, pubcrawl, q-learning, real-time modification, resilience, Resiliency, resource allocation, routing computation, routing configuration, routing optimization, routing strategies, Scalability, SDN controller, Software Defined Network (SDN), software defined networking, telecommunication control, telecommunication network routing, telecommunication traffic |
| Abstract | Software Defined Networking (SDN) provides opportunities for flexible and dynamic traffic engineering. However, in current SDN systems, routing strategies are based on traditional mechanisms which lack in real-time modification and less efficient resource utilization. To overcome these limitations, deep learning is used in this paper to improve the routing computation in SDN. This paper proposes Convolutional Deep Reinforcement Learning (CoDRL) model which is based on deep reinforcement learning agent for routing optimization in SDN to minimize the mean network delay and packet loss rate. The CoDRL model consists of Deep Deterministic Policy Gradients (DDPG) deep agent coupled with Convolution layer. The proposed model tends to automatically adapts the dynamic packet routing using network data obtained through the SDN controller, and provides the routing configuration that attempts to reduce network congestion and minimize the mean network delay. Hence, the proposed deep agent exhibits good convergence towards providing routing configurations that improves the network performance. |
| URL | https://ieeexplore.ieee.org/document/9118112 |
| DOI | 10.1109/ANTS47819.2019.9118112 |
| Citation Key | swain_codrl_2019 |
- routing configuration
- network performance
- Packet loss rate
- pubcrawl
- q-learning
- real-time modification
- resilience
- Resiliency
- resource allocation
- routing computation
- network data
- routing optimization
- routing strategies
- Scalability
- SDN controller
- Software Defined Network (SDN)
- software defined networking
- telecommunication control
- telecommunication network routing
- telecommunication traffic
- deep reinforcement learning agent
- Convolution layer
- Convolutional Deep Reinforcement Learning
- convolutional layer
- convolutional neural nets
- coupled congestion control
- current SDN systems
- Deep Deterministic Policy Gradient (DDPG)
- Deep Deterministic Policy Gradients deep agent
- deep reinforcement learning
- CoDRL model
- dynamic packet routing
- dynamic traffic engineering
- flexible traffic engineering
- gradient methods
- intelligent packet routing
- less efficient resource utilization
- mean network delay
- network congestion