KENDALI KONGESTI REAKTIF PADA JARINGAN ATM DENGAN JARINGAN SARAF TIRUAN

• Main Author: Rochim
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/1526
• Institution: Institut Teknologi Bandung

Asynchronous Transfer Mode (ATM) networks still be able to have a congested that caused by unpredictable statistical fluctuation traffic and or damaged network's component although there are ready Call Admission Controls and traffic enforcement. That way it's needed the other congestion control mechanism in manner reactive. Congestion control within ATM Networks becomes more complex, because the networks must be able serve much kind of traffic and the differ of required Quality of Service. Because of that, congestion control within ATM networks must be able adept to the change of load traffic behavior. One of adaptive computation model that could used to congestion control is Neural Networks. This research is studying, modeling, and simulation about reactive congestion control within ATM Networks by using Neural Networks that functioned to adjust the buffer threshold value and reduction the traffic rate dynamically. <br /> Through this research will be tested the hypothesis that using Neural Networks as a reactive congestion control could descend Cell Loss Ratio and how the effect to the throughput and queuing delay.