论文信息 - On Using Deep Reinforcement Learning for VNF Forwarding Graphs Placement

On Using Deep Reinforcement Learning for VNF Forwarding Graphs Placement

The placement of Virtual Network Function - Forwarding Graphs (VNF-FGs) is one of the fundamental operations in the next generation networks. However, embedding a VNF -FG is a complicated process which has been proved as an NP-hard problem. In this demo, we leverage one of the most advanced approaches in Deep Reinforcement Learning (DRL) in order to efficiently solve the VNF - FG embedding problem. In particular, we have proposed an extension of this approach in order to better explore the space of solutions while making it safer. The experimental platform we have developed, which is based on Mininet and containers, allows us to clearly show the advantage of our approach over existing approaches.

Abdelkader Outtagarts | Pham Tran Anh Quang | Yassine Hadjadj Aoul

[1] Ahmad Khonsari,et al. DeepViNE: Virtual Network Embedding with Deep Reinforcement Learning , 2019, IEEE INFOCOM 2019 - IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS).

[2] Matthew Roughan,et al. The Internet Topology Zoo , 2011, IEEE Journal on Selected Areas in Communications.

[3] Peilin Hong,et al. Virtual network function placement and resource optimization in NFV and edge computing enabled networks , 2019, Comput. Networks.

[4] Yuval Tassa,et al. Continuous control with deep reinforcement learning , 2015, ICLR.

[5] Abdelkader Outtagarts,et al. Deep Reinforcement Learning Based QoS-Aware Routing in Knowledge-Defined Networking , 2018, QSHINE.

[6] Albert Cabellos-Aparicio,et al. A Deep-Reinforcement Learning Approach for Software-Defined Networking Routing Optimization , 2017, ArXiv.

[7] Abdelkader Outtagarts,et al. A Deep Reinforcement Learning Approach for VNF Forwarding Graph Embedding , 2019, IEEE Transactions on Network and Service Management.