OPENQOS PERFORMANCE IMPROVEMENT THROUGH UTILIZING FAME ALGORITHM

• Main Author: Ginanjar - NIM : 23514057 , Bintang
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/21442
• Institution: Institut Teknologi Bandung

Qos routing is a method to find feasible path between source node and destination node which satisfy several constraints, one form of QoS routing is Delay-Constrained Least-Cost (DCLC). OpenQoS adopt DCLC to provide the Quality of Service by providing feasible path which has least cost but also has a delay that satisfy delay boundary. As a framework which uses DCLC, OpenQoS also faces several issues which usually happened in QoS routing, namely inaccurate information which has a great impact to instability routing. The instability routing can degrades the end-to-end performance. This research proposes delay estimation using Fast Algorithm for Median Estimation (FAME) as a delay estimator on OpenQoS. We employ Design Research Methodology (DRM) as research methodology. To verify the hypothesis, we build simulation environment on ns-3 network simulator and provide three different scenarios for delay distribution namely uniform, exponential and gamma distribution. According to simulation result from two different service classes, QoS Level-1 and QoS Level-2 shows that FAME algorithm as delay estimator shows better performance. For the average delay of QoS Level-1, the proposed scheme shows 68.11 ms, 59.96 ms and 127.78 ms meanwhile OpenQoS shows 69.95 ms, 61.43 ms and 131.14 ms. For the average packet loss, the proposes scheme shows 0.98%, 0.80% and 1.50% meanwhile OpenQoS shows 2.83%, 1.44% and 3.74%. <br /> <br />