Performance evaluation of packet voice multiplexer with selective packet-discarding control

Zhisheng Niu,Haruo Akimaru,Fumiaki Machihara,Ichiro Ide
DOI: https://doi.org/10.1002/ecja.4410750404
1992-01-01
Abstract:This paper evaluates the performance of packet voice multiplexer with selective packet-discarding (SPD) control. The control scheme is based on the embedded ADPCM coding technology wherein the encoded information is grouped into two separate parts: a more significant part (MSP); and a less significant part (LSP). When the buffer length exceeds a certain threshold, the LSPs of the incoming packets will be discarded to relieve the congestion. By assuming that the voice packets (cells) arrive at the multiplexer in phase-type Markov renewal process (PH-MRP) and that they are all of the fixed length, we model the system under consideration as PH-MRP(x)/-Ek/1/m(m1) with sufficiently large k. Using the matrix-analytic (MA) method, recursion formulas and effective computation algorithms for the performance measures of interest are derived. Numerical examples are illustrated.
What problem does this paper attempt to address?