TY - JOUR
T1 - SOTE
T2 - Traffic engineering in hybrid software defined networks
AU - Guo, Yingya
AU - Wang, Zhiliang
AU - Liu, Zhifeng
AU - Yin, Xia
AU - Shi, Xingang
AU - Wu, Jianping
AU - Xu, Yang
AU - Chao, H. Jonathan
N1 - Publisher Copyright:
© 2019 Elsevier B.V.
PY - 2019/5/8
Y1 - 2019/5/8
N2 - Traffic engineering (TE) is an efficient method to balance flows in a network. Existing TE solutions mainly focus on routing traffic along shortest paths between source and destination nodes, but ignore the likely network congestion. With the emergence of Software Defined Networking (SDN), through the deployment of SDN switches, we can enable flexible routing by splitting flows arbitrarily among multi-paths. However, due to technical, organizational and economical challenges of full SDN deployment, deploying SDN incrementally may be a more practical solution. Therefore, in this paper, we focus on a hybrid SDN architecture, where both legacy routers and SDN switches coexist. To explore routing optimization of TE in a hybrid SDN, we formulate a Mixed Integer Non-Linear programming (MINLP) problem and propose a heuristic algorithm SOTE to jointly optimize the OSPF weight setting and traffic splitting ratio of SDN nodes. Through extensive simulation and implementation on our prototyped testbed, we demonstrate that SOTE outperforms other algorithms in minimizing the maximum link utilization (MLU) of a network. Moreover, we also demonstrate that 10% deployment of SDN nodes can reap the most of benefit.
AB - Traffic engineering (TE) is an efficient method to balance flows in a network. Existing TE solutions mainly focus on routing traffic along shortest paths between source and destination nodes, but ignore the likely network congestion. With the emergence of Software Defined Networking (SDN), through the deployment of SDN switches, we can enable flexible routing by splitting flows arbitrarily among multi-paths. However, due to technical, organizational and economical challenges of full SDN deployment, deploying SDN incrementally may be a more practical solution. Therefore, in this paper, we focus on a hybrid SDN architecture, where both legacy routers and SDN switches coexist. To explore routing optimization of TE in a hybrid SDN, we formulate a Mixed Integer Non-Linear programming (MINLP) problem and propose a heuristic algorithm SOTE to jointly optimize the OSPF weight setting and traffic splitting ratio of SDN nodes. Through extensive simulation and implementation on our prototyped testbed, we demonstrate that SOTE outperforms other algorithms in minimizing the maximum link utilization (MLU) of a network. Moreover, we also demonstrate that 10% deployment of SDN nodes can reap the most of benefit.
KW - Hybrid SDN
KW - Maximum link utilization
KW - Open shortest path first
KW - Software defined networking
KW - Traffic engineering
UR - http://www.scopus.com/inward/record.url?scp=85062946509&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85062946509&partnerID=8YFLogxK
U2 - 10.1016/j.comnet.2019.03.008
DO - 10.1016/j.comnet.2019.03.008
M3 - Article
AN - SCOPUS:85062946509
SN - 1389-1286
VL - 154
SP - 60
EP - 72
JO - Computer Networks
JF - Computer Networks
ER -