In-service monitoring of quality of service in ATM networks using OAM cells

Abstract

Congestion conditions in ATM switches may cause gradual deterioration in the QoS. QoS monitoring, using OAM, can accurately measure the overall quality of monitored parameters to detect such deterioration before the service quality has dropped below an acceptable level. Several factors call affect the accuracy of the in-service QoS performance monitoring technique, such as the burstiness of the background traffic, network load, monitoring frequency, and bit rate. However, acceptable levels of accuracy can be achieved using OAM cells. Because of the fluctuations presented by VBR video, OAM cells can measure their characteristics more accurately than CBR video. Also, for low bursty background traffic, the cell delay variation (CDV) of VBR is higher than CBR. As the burstiness of the background traffic increases, the CDV of CBR grows faster than that of VBR. At high burstiness, the CDV of CBR is higher than VBR. For all background data burstiness, the first multiplexer of the network absorbs most of the CDV.