TY - GEN
T1 - A connectionist approach to dynamic resource management for virtualised network functions
AU - Mijumbi, Rashid
AU - Hasija, Sidhant
AU - Davy, Steven
AU - Davy, Alan
AU - Jennings, Brendan
AU - Boutaba, Raouf
N1 - Publisher Copyright:
© 2016 IFIP.
PY - 2017/1/13
Y1 - 2017/1/13
N2 - Network Functions Virtualisation (NFV) continues to gain attention as a paradigm shift in the way telecommunications services are deployed and managed. By separating Network Functions (NFs) from traditional middleboxes, NFV is expected to lead to reduced CAPEX and OPEX, and to more agile services. However, one of the main challenges to achieving these objectives is on how physical resources can be efficiently, autonomously, and dynamically allocated to Virtualised Network Functions (VNFs) whose resource requirements ebb and flow. In this paper, we propose a Graph Neural Network (GNN)-based algorithm which exploits Virtual Network Function Forwarding Graph (VNF-FG) topology information to predict future resource requirements for each Virtual Network Function Component (VNFC). The topology information of each VNFC is derived from combining its past resource utilisation as well as the modelled effect on the same from VNFCs in its neighbourhood. Our proposal has been evaluated using a deployment of a virtualised IP Multimedia Subsystem (IMS), and real VoIP traffic traces, with results showing an average prediction accuracy of 90%. Moreover, compared to a scenario where resources are allocated manually and/or statically, our proposal reduces the average number of dropped calls by at least 27% and improves call setup latency by over 29%.
AB - Network Functions Virtualisation (NFV) continues to gain attention as a paradigm shift in the way telecommunications services are deployed and managed. By separating Network Functions (NFs) from traditional middleboxes, NFV is expected to lead to reduced CAPEX and OPEX, and to more agile services. However, one of the main challenges to achieving these objectives is on how physical resources can be efficiently, autonomously, and dynamically allocated to Virtualised Network Functions (VNFs) whose resource requirements ebb and flow. In this paper, we propose a Graph Neural Network (GNN)-based algorithm which exploits Virtual Network Function Forwarding Graph (VNF-FG) topology information to predict future resource requirements for each Virtual Network Function Component (VNFC). The topology information of each VNFC is derived from combining its past resource utilisation as well as the modelled effect on the same from VNFCs in its neighbourhood. Our proposal has been evaluated using a deployment of a virtualised IP Multimedia Subsystem (IMS), and real VoIP traffic traces, with results showing an average prediction accuracy of 90%. Moreover, compared to a scenario where resources are allocated manually and/or statically, our proposal reduces the average number of dropped calls by at least 27% and improves call setup latency by over 29%.
KW - Artificial Intelligence
KW - Machine Learning
KW - Network Functions Virtualisation
KW - Neural Networks
KW - Resource Management
KW - Topology Awareness
UR - http://www.scopus.com/inward/record.url?scp=85013671828&partnerID=8YFLogxK
U2 - 10.1109/CNSM.2016.7818394
DO - 10.1109/CNSM.2016.7818394
M3 - Conference contribution
AN - SCOPUS:85013671828
T3 - 2016 12th International Conference on Network and Service Management, CNSM 2016 and Workshops, 3rd International Workshop on Management of SDN and NFV, ManSDN/NFV 2016, and International Workshop on Green ICT and Smart Networking, GISN 2016
SP - 1
EP - 9
BT - 2016 12th International Conference on Network and Service Management, CNSM 2016 and Workshops, 3rd International Workshop on Management of SDN and NFV, ManSDN/NFV 2016, and International Workshop on Green ICT and Smart Networking, GISN 2016
A2 - Keith-Marsoun, Shannon
A2 - dos Santos, Carlos Raniery Paula
A2 - Limam, Noura
A2 - Cheriet, Mohamed
A2 - Zhani, Mohamed Faten
A2 - Festor, Olivier
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 12th International Conference on Network and Service Management, CNSM 2016 and Workshops, 3rd International Workshop on Management of SDN and NFV, ManSDN/NFV 2016 and International Workshop on Green ICT and Smart Networking, GISN 2016
Y2 - 31 October 2016 through 4 November 2016
ER -