.png){kind=icon}
.png){kind=icon}
Return
A fine-grained and dynamic scaling method for service function chains
.png){kind=icon}
PRE
AI
DOI：10.1016/j.knosys.2021.107289.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Network function virtualization (NFV) is a promising approach to solve network ossification, which provides network services through virtual network function (VNF). However, the resource demands of service function chains (SFCs) frequently change during their lifecycles. Network operators need to scale the SFCs when resource demands change. The challenge of SFC scaling is that how to solve the conflict between improving the scaling success ratio and reducing the consumption of resources. This paper proposes a hybrid scaling method (HSM) that effectively solves this conflict. First, we formulate the SFC scaling problem as an integer linear programming model. We then present the improved vertical scaling (IVS) and horizontal scaling (IHS). IVS improves the scaling success ratio by combining vertical scaling and traffic splitting. IHS reduces CPU, forwarding and memory resource consumption by providing resources for VNF instances according to resource demands. IHS considers hop constraints when selecting server nodes and substrate links to deploy new instances and virtual links, thereby reducing the bandwidth resource consumption. The HSM further improves the scaling success ratio and reduces the scaling resource consumption by combining the IVS and IHS. Finally, through theoretical analysis and simulation experiments, we demonstrate the effectiveness of the proposed hybrid scaling method. (C) 2021 Elsevier B.V. All rights reserved.
Keywords:
Service function chain
Network function virtualization
Virtual network function
Dynamic scaling
Fine-grained resource provision
Journal
K
IF:
7.6
Papers:
1.2W
Citations:
4.5W