A Fuzzy-Based Dynamic Bandwidth Allocation Approach for Campus Area Networks

A robust campus area network (CAN) with excellent connectivity to the internet is no longer a luxury to the educational institutions since the internet plays a vital role in education. With that, proper bandwidth allocation is needed so that the internet will be utilized for educational purposes. This study proposes a fuzzy based dynamic bandwidth allocation approach which can be used in CAN. The Mamdani fuzzy inference system was utilized along with the enhanced web usage mining (WUM) algorithm to achieve the proposed method. The improved WUM was employed to the weblog, captive portal log, and Shallalist file of the proxy server to identify the network users and the websites they accessed together with the website categories. Whereas the Mamdani FIS was used to decide the bandwidth that will be allocated to the users based on the rules provided. A prototype was developed and installed in a pfSense proxy server. The prototype was tested with several users connected to the proxy. The results showed that the prototype could dynamically allocate bandwidth to the computers and the bandwidth is based on the given rules. Further, this method is vital to the network administrators since the setting of the bandwidth can be done automatically.