Open AccessRTSM: Response time optimisation during switch migration in software‐defined wide area network
Author(s) -
Sahoo Kshira Sagar,
Tiwary Mayank,
Sahoo Bibhudatta,
Mishra Brojo Kishore,
RamaSubbaReddy Somula,
Luhach Ashish Kr.
Publication year - 2020
Publication title -
iet wireless sensor systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.433
H-Index - 27
ISSN - 2043-6394
DOI - 10.1049/iet-wss.2019.0125
Subject(s) - computer science , controller (irrigation) , quality of service , scalability , response time , software defined networking , reliability (semiconductor) , forwarding plane , can bus , process (computing) , real time computing , distributed computing , computer network , operating system , power (physics) , physics , quantum mechanics , network packet , agronomy , biology
The distributed control plane is the alternate solution to reliability and scalability like potential issues in the software‐defined wide area network (SDWAN), but the static mapping between controller and switches might cause an uneven load distribution among controllers. Migration of switches from one controller to another under‐loaded controller is a solution to handle the peak traffic. However, the switch migration process is a complex process that may affect the end‐user quality of service (QoS), as a result, the response time of the control messages are also affected. To address this issue, the authors present RTSM, a novel strategy, that optimises the response time of the control messages during switch migration. Further, an SDWAN architecture was proposed, in which the introduction of the layer‐2 controller reduces the dependency on WAN communication of the forwarding devices. The Karush–Kuhn–Tucker conditions have been applied for target controller selection, which ensures improved response time. A switch selection method was also introduced, which minimally affects the end‐user QoS during device migration in SDWAN scenario. To evaluate the performance of the RTSM, Mininet simulator with the Floodlight controller has been used, and the result shows that the proposed algorithm outperforms the other existing works during load balancing.