Performance evaluation of AAL2 voice multiplexer

This paper investigates the traffic performance of an ATM Adaptation Layer 2 (AAL2) cell assembly and disassembly device (CLAD), which is an AAL2 multiplexer, by an analytical method where AAL2 is standardized to convey low-bit-rate traffic such as mobile traffic including IMT-2000. In particular, it evaluates the number of channels that can be multiplexed in the fixed capacity output link of the AAL2 multiplexer for various timeouts and buffer sizes. The arrival process, modeled as a discrete time Markovian arrival process accounting for the talkspurt and silence period of a voice source, and the cell assembly queue and transmission queue, modeled as discrete time tandem queues are analyzed through an embedded Markov-chain technique. Approximation models and their accuracy are also investigated and the Independence approximation is shown to be good in terms of accuracy and computation time. The short packet loss ratio, which is lower than the cell loss ratio, is derived. Through numerical examples, this paper shows the statistical multiplexing gain of AAL2 and the various parameter ranges that yield higher efficiency than STM. It also shows that timeouts have a complicated impact on the performance of AAL2, as a result of changes in the characteristics of cell arrival at the transmission queue. The scheme, in which multiple CLADs are accommodated in a virtual path, is also investigated.