Isyaku, Babangida (2017) Analyzing performance of openstate in software defined network with multiple failures scenarios. Masters thesis, Universiti Teknologi Malaysia, Faculty of Computing.
|
PDF
464kB |
Official URL: http://dms.library.utm.my:8080/vital/access/manage...
Abstract
Software Defined Network (SDN) is an emerging network that decouples the control plane and data planes. Like other networks, SDN undergoes a recovery process upon occurrences of link or node failures. Openflow is considered as the popular standard used in SDN. In Openflow, the process of detecting the failure and communications with controller to recompute alternative path result to long recovery time. However, there is limit with regards time taken to recover from the failures. If it takes more than 50 msec, a lot of packet will be lost, and communication overhead and Round Trip Time (RTT) between switch – controller may be high. Openstate is an Openflow extension that allows a programmer to specify how forwarding rules should be adapted in a stateful fashion. Openstate has been tested only on single failure. This research conduct experiment based on Openstate pipeline design that provides detections mechanism based on switches periodic link probing and fast reroute of traffic flow even when controller is not reachable. In this research, the experiments use Mininet simulation software to analyse and evaluate the performance of Openstate with multiple failure scenarios. The research has compared Overhead communication, Round Trip Time (RTT) between switch – controller and number of packet loss with Openflow and Openstate. On the average, in Openstate packet loss is zero when the recovery time is less than or equal to 70 msec while communication overhead involves 60 packet-in. In Openflow, packet loss is zero when the recovery time is less than or equal to 85 msec while communication overhead involves 100 packet-in. Finally, the average RTTs for Openstate and Openflow are 65 msec and 90 msec respectively. Based on the results obtained, it can be concluded that Openstate has better performance compare to Openflow.
Item Type: | Thesis (Masters) |
---|---|
Additional Information: | Thesis (Sarjana (Sains Komputer)) - Universiti Teknologi Malaysia, 2017; Supervisor : Assoc. Prof. Dr. Mohd. Soperi Mohd. Zahid |
Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
Divisions: | Computing |
ID Code: | 78536 |
Deposited By: | Fazli Masari |
Deposited On: | 27 Aug 2018 03:22 |
Last Modified: | 27 Aug 2018 03:22 |
Repository Staff Only: item control page