4142

3912

كارشناسي ارشد

مخابرات

اصفهان : داتشگاه صنعتي اصفهان : دانشكده برق و كامپيوتر

1378

بيست ، 118 ، [II] ص . : مصور ، جدول ، نمودار

ص . ع . به : فارسي و انگليسي

حسين سعيدي

محمد رضا هاشمي

انگليسي به فارسي

20/7/1378

مهندسي برق و كامپيوتر

ID3912

به فارسي و انگليسي : قابل رويت در نسخه ديجيتال

Abstract With the advent of packet based networks and new applications in such networks efficient use of network resources congestion control and fair allocation of bandwidthbecome very important Traffic scheduling algorithms provide a mean to reach these goals Such algorithms determine the service order of packets of different sessions which aresharing an outgoing link Two main objectives in designing scheduling algorithms are toreach fairness in bandwidth allocation and to reduce the worst case delay of packets Inaddition time complexity of the algorithms is also a very important factor especially inbroadband networks The service time of a packet is very short in these networks and thealgorithm must complete its run in such a short period Hence designing low complexityalgorithms with desirable fairness properties and limited worst case delay of packets is thegoal This thesis tries to categorize different scheduling algorithms and study the propertiesof each category Sorted priority based scheduling algorithms as a main and importantclass of traffic scheduling algorithms is concentrated on and algorithms of this category arestudied and compared in terms of time complexity fairness and delay Golestani s fairnessindex is used to compare their fairness properties and LR servers theory is invoked tocompare worst case delay of packets in these algorithms Two main approaches areintroduced to improve current algorithms one is based on achieving lower fairness indexand the other is based on reaching lower worst cast delay for packets According to theseapproaches two algorithms called LVT SCFQ and FPFQ are introduced and theirperformances are evaluated in terms of fairness delay and complexity

حسين سعيدي

محمد رضا هاشمي