Premium
SDN/NFV‐enabled performance estimation framework for SFC optimization
Author(s) -
Yi Bo,
Wang Xingwei,
Huang Min
Publication year - 2018
Publication title -
transactions on emerging telecommunications technologies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.366
H-Index - 47
ISSN - 2161-3915
DOI - 10.1002/ett.3441
Subject(s) - computer science , chaining , virtual network , distributed computing , virtualization , agile software development , software deployment , orchestration , service (business) , cloud computing , software engineering , operating system , psychology , musical , art , economy , economics , visual arts , psychotherapist
Many network function virtualization management and orchestration frameworks have been proposed to offer an agile and automated service deployment with a set of virtual network function (VNF) placement and chaining algorithms. However, once the services are deployed, less attention is paid to the subsequent supervision on their performance which may be affected by the dynamic changing network conditions. In order to guarantee the performance of services, this article designs and presents a service function chain performance estimation framework, which can be used to complement the network function virtualization orchestrators by providing performance estimation and optimization for the already deployed services. In particular, the performance estimation is implemented based on the min‐plus algebra theory, and the optimization is fulfilled by using two VNF migration approaches on the basis of greedy and load balance, respectively. The proposed framework can (1) evaluate and compare existing VNF placement algorithms by statistics on service performance estimation, (2) estimate and optimize the performance of services by using VNF migration approaches, (3) integrate the VNF placement algorithms and VNF migration approaches, thus producing more efficient future solutions. Finally, this article presents an implementation of the proposed framework and evaluates its effectiveness in an experimental environment.