An SDN-based slow start algorithm for data center networks

With the fast development of Cloud Computing and online business of Internet, the importance of data center is gradually increasing. Because of the high bandwidth and low latency characteristics of the data center network (DCN), the tradition transmission control protocol (TCP) slow start degrades the performance of the data center. In this paper, based on the software-defined networking (SDN) technology, we propose a new TCP slow start algorithm for DCNs, named the soft-defined slow start (SDSS) algorithm. In the proposed SDSS algorithm, the available bandwidth can be obtained by the SDN controller in every time slot, that the slow start threshold (ssthresh) and the congestion window (cwnd) can be set appropriately. Because the data transmission rate does not exceed the available bandwidth in every time slot, the risk of congestion can be minimized. Simulation results show that the proposed SDSS algorithm can increase the throughput and avoid the congestion of the network compared with the existing algorithms, especially for the mice flow.

[1]  Carey L. Williamson,et al.  A new scheme for TCP congestion control: smooth-start and dynamic recovery , 1998, Proceedings. Sixth International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (Cat. No.98TB100247).

[2]  Ken-ichi Kitayama,et al.  Modeling and performance analysis of OPS data center network with flow management using Express Path , 2014, 2014 International Conference on Optical Network Design and Modeling.

[3]  Nick McKeown,et al.  OpenFlow: enabling innovation in campus networks , 2008, CCRV.

[4]  Giovanni Pau,et al.  TCP Start up Performance in Large Bandwidth Delay Networks. , 2004, INFOCOM 2004.

[5]  Peter Steenkiste,et al.  Improving TCP startup performance using active measurements: algorithm and evaluation , 2003, 11th IEEE International Conference on Network Protocols, 2003. Proceedings..

[6]  S. Ha,et al.  Hybrid Slow Start for High-Bandwidth and Long-Distance Networks , 2008 .