Return to search

A Dynamic Throughput Improvement Scheme with Priority Queues in Differentiated Services Networks

Differentiated-Service networks is designed for solving scalability problems through traffic aggregation. However, it can't guarantee end-to-end QoS of individual flow. In this thesis, we propose a Self-Adaptive Control Scheme for Differentiated-Service networks that can improve the throughput of individual flows dynamically. In this scheme, egress routers monitor the average throughput of individual flow, and send the Self-Adaptive Control Messages to ingress routers if need. The ingress router re-allocate network resources to improve throughput of high-priority flows depending on the Control Messages.
We use NS-2 simulator to prove that our scheme that can improve throughput of high-priority flows dynamically, and suggest that a better time interval of Self-Adaptive control can be determined based on the queue sizes, packets arrival rate and departure rate. Finally, we use Random Early Detection (RED) queue instead of Drop-Tail queue to reduce unfairness of individual flows when there are congestion and insufficient network resources.

Identiferoai:union.ndltd.org:NSYSU/oai:NSYSU:etd-0726100-172436
Date26 July 2000
CreatorsTseng, Fan-Geng
ContributorsWen-Shyen Chen, Cheng-Shong Wu, Tsang-Ling Sheu, Wei-Kuang Lai, Kim-Joan Chen
PublisherNSYSU
Source SetsNSYSU Electronic Thesis and Dissertation Archive
LanguageCholon
Detected LanguageEnglish
Typetext
Formatapplication/pdf
Sourcehttp://etd.lib.nsysu.edu.tw/ETD-db/ETD-search/view_etd?URN=etd-0726100-172436
Rightsunrestricted, Copyright information available at source archive

Page generated in 0.0021 seconds