Load Balancing with a New Approach for Dynamic Server Selection in Software Defined Network
The 1st International Conference on Electrical-Electronics and Computer Engineering (ICEECE 2021), Trabzon, Turkey, 24 - 25 July 2021, (Full Text)
- Publication Type: Conference Paper / Full Text
- City: Trabzon
- Country: Turkey
- Karadeniz Technical University Affiliated: Yes
Abstract
Today, fulfilling the increasing internet demands with a single server is a very laborious task. To overcome this problem, more than one server must be used depending on the density of client requests. Load balancing in Software Defined Networks is defined as the distribution of the load between different servers depending on the load balancing strategy determined. While applying the load balancing strategy, it is aimed to respond to the incoming requests as soon as possible by the selected server. Routing with forecasting is very difficult as the traffic flow is constantly changing. Therefore, when the flow comes, the choice decision must be made immediately. While making the decision, statistical information is collected periodically from the switch ports to which the servers are connected and the server is selected according to the determined decision function. The most important step that determines the success of the load strategy is to determine the parameters used in the decision function and to what extent these parameters contribute to the function. The decision function used for server selection in this publication is originally designed. The server is dynamically selected according to the decision function designed to adapt to the chancing traffic flow. POX, which supports Python programming language, is preferred as a controller. Mininet Emulator is used to create the network topology. The proposed new approach for server selection is compared with traditional server selection methods. And the results prove that the designed decision function reduces the response time and increases the throughput.