DocumentCode
2244913
Title
Multicast-based inference of network-internal delay performance using the method of moment
Author
Lin, Junwu ; Zhang, Jianzhong ; Lin, Wen
Author_Institution
Dept. of Electron. Eng., Xiamen Univ., Xiamen, China
Volume
2
fYear
2010
fDate
6-7 March 2010
Firstpage
193
Lastpage
196
Abstract
Packet delay has an important influence on the overall performance of network in applications, therefore, it is very necessary to measure internal-network delay performance for network administrators to manage their network well. Existing methods, largely based on end-to-end measurement with discrete delay model, can be applied to infer delay performance of internal links, such as Maximum Likelihood Estimate (MLE) and EM-MLE (Expectation Maximum) algorithm, however, their computational complexity caused by the iterative inference is still a problem for practical use, especially, it is more complex in computation when the number of bins is more or the size of network topology is larger. In this paper, we show how to infer the delay performance of internal network using the method of moment. In contrast to other methods, its most advantage is simple in computational complexity, and for the large network topology, it has more advantage in computational complexity than other methods.
Keywords
Internet; expectation-maximisation algorithm; inference mechanisms; maximum likelihood estimation; multicast communication; discrete delay model; expectation maximization algorithm; maximum likelihood estimation; method-of-moment; multicast-based inference; network-internal delay performance; packet delay; Computational complexity; Computer network management; Delay estimation; Inference algorithms; Iterative algorithms; Iterative methods; Maximum likelihood estimation; Moment methods; Network topology; Size measurement; delay; moment method; multicast; network tomography;
fLanguage
English
Publisher
ieee
Conference_Titel
Informatics in Control, Automation and Robotics (CAR), 2010 2nd International Asia Conference on
Conference_Location
Wuhan
ISSN
1948-3414
Print_ISBN
978-1-4244-5192-0
Electronic_ISBN
1948-3414
Type
conf
DOI
10.1109/CAR.2010.5456571
Filename
5456571
Link To Document