Manajemen Bandwidth Berdasarkan Prediksi Perilaku Pengguna Pada Jaringan TCP/IP Dengan Jaringan Syaraf Tiruan

Abstract

User behavior prediction on TCP / IP networks is very possible because each data packet is passed through a router that can be analyzed and then classified into several behavior types from commonly performed activities by internet users. User-behavior predictions developed using artificial neural networks (ANN), which can show accurate results in predicting the bandwidth used by users in the next few days. Predicting results helps network administrators to provide a more accurate picture of user behavior patterns in the future, where the Personnel Service Application System (SAPK) which becomes the user behavior internet priority it's only utilized network resources at 26.8% of the total internet network resources available. The ANN that was developed also showed that 40.3% of network resources were used more by streaming video activities and 37% were used for streaming audio activities. With the prediction result, the network administrator rearranges the internet network resources distribution by applying a new prediction results pattern. This study conclusively shows that user behavior with streaming video activities is the largest bandwidth user and needs to get special attention on the internet network at the Regional X National Civil Service Agency Office.