Profiling network traffic of Sultan Idris Shah building (BSIS) using data mining technique / Rusmawati Ishak

With the exponential of information technology and data growth, a rapid increase in the request to the system and workload may cause severe congestion. A high performance of BSIS network infrastructure and services that is capable to serve thousands or millions of traffic for BSIS users faster, more reliable, lower response time and will bring good experience to the online users. Hence, in order to make it reliable to provide good online services, we need to identify what kind of application, characteristic trend and behaviour of BSIS users in network infrastructure. Thus, in this research we focus on analysing BSIS network traffic to find a pattern that cause of slow connection during peak hours and non-peak hours that will come out with network profiling information. The network profiling will help the agency to give better solution in managing network and security policy and infrastructure as well as to give better experienced in online activities request by users. This research implement unsupervised data mining approach to analyse the network traffic trend of BSIS. Orange is a tool that being used in implementing K-Means Clustering Algorithm that could be seen as the most suitable solution to find a network trend pattern of user accessing the Internet and to produce with profiling network. The result showed that the K-Means method can perform clustering with three (3) cluster in which describe based on high, medium and low number of hits towards the protocol services and unique IP address. Based on the result, most higher traffic pass through the network that consume of bandwidth usage are DNS (UDP) with total of 74,819 hits, MySQL with total of 40,691 and HTTP service with 59,784 hits record. Based on the trend shows more activities done during the holiday and the attacking activity also frequently happen during the weekend or public holiday. Data mining process lead to reveal the information gather for profiling purposes and identify type of traffic passing through the BSIS network. Furthermore, the outcome of this study can be a recommendation of managing or shaping the bandwidth usage and strengthen the security policy as well as help the organization to manage the network and assist the connectivity issue faced by internal and online user.