Measurement-based delay performance estimation in ATM networks

Real-time traffic requires strict delay constraints in terms of maximum cell transfer delay(CTD) and cell delay variation(CDV). In order to support the required quality of service (QoS) of each connection in ATM networks, it is important to estimate the current status of networks by using local information such as cell delay and delay variation in switches. We develop a framework for obtaining delay distribution in a single switch module by monitoring the internal buffer. We also propose a new delay estimation mechanism utilizing a fast convolution approximation scheme in order to reduce the number of operations and the required amount of the data for estimating end-to-end delay. The feasibility of the proposed delay estimation mechanism is verified by simulation for various types of CBR and VBR traffic loads.