Controlling Traffic Congestion on Wireless Networks: A Multinomial Distribution Approach
| Olaniyan Olatayo Moses Bells University of Technology, Nigeria This e-mail address is being protected from spambots. You need JavaScript enabled to view it ) | Shakirat Oluwatosin Raji Bells University of Technology, Nigeria ( This e-mail address is being protected from spambots. You need JavaScript enabled to view it |
Congestion is the degradation in performance of network when the subnet is dumped with too much traffic which usually occurs at the bottleneck nodes. Since Wireless networks have become essential telecommunication infrastructure as millions of people depend on these networks for daily communication Therefore, it is often required to address the problems linked with network congestion so as to get the Quality of Service(QoS) required by end –users. The purpose of this research is to describe the behavior of the users and packet traffic flowing on realistic networks as packet load is varied on the University Networks and also to investigate the causes and effects of traffic congestion on the University networks. The purpose of congestion control is to ensure network stability and achieve a reasonably fair distribution of the network resources among the users.
The growing popularity of wireless networks has led to cases of heavy utilization and congestion. Some researchers Baker et al (1999) have suggested that the burstiness in packet traffic is connected to the behavior of individual users within the network. However, the modeling of packet traffic is based on its measured characteristics more than on the underlying mechanisms responsible for the self-similarity. This paper uses the data information collected from Bell University of Technology Ota networks, on different nodes to discover total numbers of user on each node at a particular time and total packets sent, and then the data was used to generate the model. This paper uses WLAN in Bells University to show the adverse effects of congestion on a network. The whole network has four access points nodes spread across the institution. Additionally, new users are joining the wireless service every day. As the networks grow, users face tremendous challenges (congestion) since congestion of wireless network might change as it expands over time. This research uses Mathematical model to determine the behaviour of the users flowing on the networks and the level of congestion for different periods of time for each day. This model is quite general and can be applied to any networks. This model is formulated using a Multinomial Distribution.
Object oriented programming language (JAVA) was used to simulate the whole networks. It shows the level of congestion of each node and how to control the congestion levels using progress and comment bars. Chi square was also used as a mathematical tool to know the significance difference of the total number of users at any period of the time. The most congested node in the whole networks was identified and the time for heavily, slightly, moderately and bad congestion period was inferred from the investigation.
This research is to understand the congestion condition and provide the solution useful to network administrator and operators of large network infrastructures. Such understanding offers network administrator valuable insights about predicting the frequency or probability with which network congestion occurs and provides solution to the problem. Further work can be done to investigate more on the effect and causes of congestion on the networks and how to avoid congestion.
Olaniyan Olatayo Moses: Lecturer at the Department of Computer Science Bells University of Technology Ota, Ogun State. M.Sc. Computer Science from University of Ibadan 2009.B.Tech Computer Engineering from Ladoke Akintola University Of Technology, Ogbomoso 2005.
Area of research interest: Networking, Medical Information System and Human Computer Interaction.
