Synchronization algorithm for stream control transmission protocol (SCTP) network.
Item
-
Title
-
Synchronization algorithm for stream control transmission protocol (SCTP) network.
-
Identifier
-
AAI3103107
-
identifier
-
3103107
-
Creator
-
Elsayed, Hussein A.
-
Contributor
-
Adviser: Tarek Saadawi
-
Date
-
2003
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Engineering, Electronics and Electrical
-
Abstract
-
Synchronization has been a problem in communication networks especially for real time applications. Synchronization can be done in different layers of the ISO model. In this Thesis, a synchronization algorithm is introduced and applied on Stream Control Transmission Protocol (SCTP), which is a layer four protocol. However, the current SCTP standard, RFC 2960, doesn't specify any synchronization mechanisms. We examine the Transfer Delay (TD) and the Interarrival Time (IT) as network performance measure. The Algorithm is applied on periodic and non-periodic traffic and shows significant synchronization improvement.;The periodic traffic algorithm has three levels depending on the reference delay, namely, "Maximum Delay Algorithm", "Accumulative Average Algorithm", and "Average Over n Algorithm". The Maximum Delay Algorithm adjusts the jitter based on the maximum delay of the preceding packets. The Accumulative Average Algorithm keeps track of the average delay and considers it as the reference delay to adjust the upcoming packets. Finally, the Average Over n Algorithm takes the average over n preceding packets as the reference for jitter adjustment. Our results show good deal of jitter improvement. It also shows that the improvement depends on the input traffic distribution and the amount of its jitter.;This Thesis also provides an analytical model for the periodic traffic synchronization algorithm. The analysis uses a basic principle in probability, which is called "transformation of random variables". Since the analytical model depends on the probability distribution function, we analyze the uniform and normal distribution to show how much improvement our algorithm provides. Other distribution can be analyzed but they come with similar conclusion. We also found upper and lower limits for the mean and variance of the output traffic delay independent of the distribution.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
