Header menu link for other important links
X
Jitter buffer modelling and analysis for TDM over PSN
, Rathinam M.
Published in Elsevier B.V.
2015
Volume: 62
   
Pages: 186 - 195
Abstract
Time Division Multiplexing (TDM) over Packet Switched Network (PSN) is a pseudo wire technology for emulating TDM Circuits over Packet Networks. Conceptually, the important ingredients of the above technology are to implement the following, in the PSN (i) Quality of Service (QoS) which is implemented through scheduling at the intermediate nodes that gives priority to packets containing 'TDM' payload (ii) timing and synchronization and (iii) scheduling in the jitter buffer for minimum output variance. Among these, this paper addresses (iii) as a first step assuming that the PSN provisions the "unacknowledged virtual circuit" (the main components of 'virtual circuit' are QoS and connection-oriented service). This work targets to implement a scheduling algorithm (service intervals) in jitter buffer at the receiver, such that the variance of inter-departure intervals of TDM stream is minimized. This is accomplished by the buffer modelled as M/G/l queueing system with Auto-Regressive AR (1) correlations within service intervals. The motivation for the above correlation structure is two-fold. First, given the correlations within the service intervals, such a correlation results in reduction of variance in the inter-departure interval. The other is that the analysis of such a correlated queue is analytically tractable. The variance of the inter-departure time is presented. The analysis of the departure process, the waiting times of incoming packets of this correlated queue aids in determining the correlation parameter that are sub-optimal in the context of TDMoPSN. Our study also includes a M/G/l queue with AR (1) cross-correlations between the inter-arrival and the service times. A G/G/l queue in which the inter arrivals are correlated, and with AR (1) correlations of the above two types are also studied. Extensive simulations demonstrate our analytical and approximation results. © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license.
About the journal
JournalData powered by TypesetProcedia Computer Science
PublisherData powered by TypesetElsevier B.V.
ISSN18770509
Open AccessNo