On the effects of the packet size distribution on FEC performance

Abstract

For multimedia traffic like VBR video, knowledge of the average loss probability is not sufficient to determine the impact of loss on the perceived visual quality and on the possible ways of improving it, for example by forward error correction (FEC) and error concealment. In this paper we investigate how the packet size distribution affects the packet loss process, i.e. the probability of consecutive losses and the distribution of the number of packets lost in a block of packets and the related FEC performance. We present an exact mathematical model for the loss process of an MMPP + MMPP/E r /1/K queue and compare the results of the model to simulations performed with various other packet size distributions (PSDs), among others, the measured PSD from an Internet backbone. The results show that analytical models of the PSD matching the first three moments (mean,variance and skewness) of the empirical PSD can be used to evaluate the performance of FEC in real networks. We conclude that the exponential PSD, though it is not a worst case scenario, is a good approximation for the PSD of today's Internet to evaluate FEC performance. We also conclude that the packet size distribution affects the packet loss process and thus the efficiency of FEC mainly in access networks where a single multimedia stream might affect the multiplexing behavior. We evaluate how the PSD affects the accuracy of the widely used Gilbert model to calculate FEC performance and conclude that the Gilbert model can capture loss correlations better if the CoV of the PSD is high.

Performance evaluation of an optical transparent access tier based on PON and spectral codes

“…It has been observed in previous measurement studies that the packet size distribution typically follows a bimodal distribution, that can be accurately characterized by the following combination of two Gaussian distributions (see [10]):…”

“…The values used in [10] to empirically characterize the packet size distribution are α 1 = 0.74, μ 1 = 127 bytes, α 2 = 0.26, μ 2 = 1366 bytes, and σ 1 = σ 2 = 20 bytes. In this case:…”

“…In all packet size distributions, the mean value has been assumed EX = 0.74 × 127 + 0.26 × 1366 = 449.14 bytes, since this is the average empirical packet size [10]. This gives an average transmission delay of 23.14 μs at B = 155.52 Mbps.…”

“…Unlike the authors in [9], this work proposes the use of orthogonal codes that do not employ time-domain chips and we further analyze the coherence of state from a packet level. The analysis also assumes different packet size distributions, and a bimodal distribution that accurately captures the measured packet size distribution in real networks [10]. Finally, coherence of state has a direct application in the definition of medium access control (MAC) protocols that control the access to the passive optical network.…”

Packet loss characteristics for M/G/1/N queueing systems

“…In (Dán et al 2006a) transmission times are either deterministic or exponentially distributed and packets of the tagged flow arrive in accordance with a Poisson process. In (Dán et al 2006b), transmission times are Erlang distributed and the tagged flow is modelled by means of a Markov modulated Poisson process.…”