Title :
Link capacity estimation in wireless software defined networks
Author :
Farzaneh Pakzad;Marius Portmann;Jared Hayward
Author_Institution :
School of ITEE, The University of Queensland, Brisbane, Australia
Abstract :
SDN is a new approach to manage networks with a centralised, global view and control of the network, and a more fine grained and flexible approach to routing and forwarding of data packets. This has shown to achieve significantly increased network efficiency in a range of wired networks. SDN also has a great potential for wireless networks. One of the unique challenges of applying the SDN approach to wireless networks, in contrast to wired networks, is the dynamic nature of wireless links and the uncertainty about their capacity. In order to be able to do optimal routing and traffic engineering with SDN in a wireless network, it is critical to know the capacity of the available wireless links. This paper presents a link capacity estimation mechanism that can be implemented on any OpenFlow SDN controller. For this, we adapted the well-known technique of packet pair/train probing, and developed a method that allows the correction of estimation errors induced by cross traffic. We have implemented a prototype for the Ryu SDN controller, and our emulation-based experimental results show a promising accuracy of our proposed approach.
Keywords :
"Estimation","Control systems","Wireless communication","Dispersion","Probes","Delays","Bandwidth"
Conference_Titel :
Telecommunication Networks and Applications Conference (ITNAC), 2015 International
DOI :
10.1109/ATNAC.2015.7366814