Performance Enhancement of Floodlight Software Defined Networking Controller using Workload Adaptive Packet Batching
| dc.contributor.advisor | Yihenew, Wondie (PhD) | |
| dc.contributor.author | Petros, Belachew | |
| dc.date.accessioned | 2019-10-21T08:28:03Z | |
| dc.date.accessioned | 2023-11-04T15:14:41Z | |
| dc.date.available | 2019-10-21T08:28:03Z | |
| dc.date.available | 2023-11-04T15:14:41Z | |
| dc.date.issued | 2019-07 | |
| dc.description.abstract | The innovation of high tech devices with increasing demand for big data processing, made the networking systems unresponsive to the need of users. The capacity of network technologies such as wired, wireless, and cellular networks has been increasing highly due to the high traffic of the network system. Such traffic nature of today’s network system is very complex which is very hard to be handled by the conventional networks. These conventional network characteristics could not be adapted to the fluctuating requirements. Due to this, it is very hard to manage the different networking devices; inflexibility to increase in size of the networks, and dependability to the specific vendor's software. Software Defined Networking was made by separating the cumbersome network control from data forwarding devices, apart from traditional networking. SDN was aimed at making a networking paradigm that responds quickly to the changing network requirements. SDN controller uses an OpenFlow protocol, which handles rules for the traffic that arrives at the switch. Floodlight controller uses static packet batching for supervising the traffic by OpenFlow protocol. The static batching is sluggish to the rapidly expanding traffic and it takes a high time for processing. In this thesis, the Workload Adaptive Packet Batching, which learns the batch size based on the nature of the workloads, was proposed to optimize the performance of the Floodlight SDN controller. This study implemented and tested on the network that was created by Mininet network emulator. The network consists of the virtual switches and hosts that are managed by the Floodlight controller. After network setup, the performance evaluation was performed using the Cbench tool, which tests for throughput and latency metrics. The proposed Workload Adaptive Packet Batching achieved an enhanced average throughput of 11% and a latency of 10%. The throughput was improved and the latency was reduced with this proposed mechanism. Therefore, enterprise SDN networks can boost their performance and traffic management by applying the Floodlight controller into their networks. | en_US |
| dc.identifier.uri | http://etd.aau.edu.et/handle/123456789/19542 | |
| dc.language.iso | en_US | en_US |
| dc.publisher | Addis Ababa University | en_US |
| dc.subject | Big data | en_US |
| dc.subject | Software Defined Networking (SDN) | en_US |
| dc.subject | OpenFlow | en_US |
| dc.subject | Floodlight controller | en_US |
| dc.subject | Workload Adaptive Packet Batching | en_US |
| dc.subject | Cbench | en_US |
| dc.title | Performance Enhancement of Floodlight Software Defined Networking Controller using Workload Adaptive Packet Batching | en_US |
| dc.type | Thesis | en_US |