Future B-ISDN networks, based on asynchronous transfer mode (ATM), will be able to offer services with statistically guaranteed quality of service (QoS). A typical target is a very low cell loss probability, which raises the problem of rare events probability estimation. Moreover, experimental measurements highlighted the bursty nature of real data paths, which cannot be captured by traditional Markovian models, while this behaviour is web described by self-similar processes. To cope with these two issues, in this work we applied importance sampling (IS) techniques to the simulation of realistic broadband traffic, In particular we pointed out the effectiveness of a heuristic IS approach in evaluating the performances of a single server queueing system, loaded by realistic traffic traces, that is taking into account the bursty behaviour of real data paths.
Efficiency of Importance Sampling Technique for Queueing Performance Evaluation with LRD Traffic
GIORDANO, STEFANO;PAGANO, MICHELE;RUSSO, FRANCO;
1998-01-01
Abstract
Future B-ISDN networks, based on asynchronous transfer mode (ATM), will be able to offer services with statistically guaranteed quality of service (QoS). A typical target is a very low cell loss probability, which raises the problem of rare events probability estimation. Moreover, experimental measurements highlighted the bursty nature of real data paths, which cannot be captured by traditional Markovian models, while this behaviour is web described by self-similar processes. To cope with these two issues, in this work we applied importance sampling (IS) techniques to the simulation of realistic broadband traffic, In particular we pointed out the effectiveness of a heuristic IS approach in evaluating the performances of a single server queueing system, loaded by realistic traffic traces, that is taking into account the bursty behaviour of real data paths.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
