No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Admission Control for Speech Traffic in the Presence of Overbooking
Abstract
Voice codecs use silence detection to reduce their average output rate and produce on/off packet streams. When several of such on/off packet streams are multiplexed onto a single link, the packets face both packet scale and burst scale congestion. If the sum of the peak rates of the flows does not exceed the link bandwidth, the distribution of the packet waiting time resulting from packet scale queuing can be calculated by a weighted n·D/D/1 formula [1, 2]. However, advantage can be taken of the reduced flow rates by overbooking the link bandwidth in the sense that the sum of their mean rates is below the link bandwidth, but the sum of their peak rates exceeds the link bandwidth. Then, the well-known formula cannot be applied, but this is the most frequent application scenario. Therefore, we change the weighted formula by a simple adaptation and show by means of simulation that it well approximates the distribution of the packet waiting times in an overbooked system. We also clarify the relation of our new method to the well-known AMS [3] solution for on/off fluid flows. The simplicity of our approximation makes it attractive for engineers and applicable for admission control purposes in switching devices.
... In particular, the method works well if the traffic has the on/off characteristics of typical compressed speech [23]. The presented methods are useful in an environment with overbooked transmission lines to support admission decisions for new flows in order to avoid extensive packet loss and delay [24]. ...
We present simple approximation formulae for the distribution of the packet waiting time of multiplexed periodic traffic. The multiplexed streams may have different periods and packet sizes. We show by extensive simulations the accuracy of the proposed methods. They are simpler than other existing formulae which make them attractive for engineers, applicable in practice, and easy to implement in switching devices. Packetized speech traffic has a periodic flow structure. Many compression techniques preserve it during talk phases but suppress the generation of packets during silence phases. When such on/off streams are multiplexed, advantage can be taken of their reduced flow rates by overbooking the link bandwidth. We adapt the proposed formulae to cope with on/off traffic and overbooking and validate them by extensive simulations. They can be applied for admission control in networks carrying different types of real-time traffic. Index Terms—waiting time distribution; on/off streams; mul-tiplexing; overbooking.
A physical model is considered in which a buffer receives messages from a finite number of statistically independent and identical information sources which asynchronously alternate between exponentially distributed periods in the on and off states. While on, a source transmits at a uniform rate. The buffer depletes through an output channel with a given maximum rate of transmission. The equilibrium buffer content distribution is analyzed. The above is a useful model for a data handling switch in a computer network. A numerically stable scheme is presented for solving the differential equations. This scheme's effectiveness depends on the efficiency of the computation of all the system's eigenvalues. A simple expression is given for the asymptotic behavior of buffer content. Numerical results are presented. Bibtex entry for this abstract Preferred format for this abstract (see Preferences) Find Similar Abstracts: Use: Authors Title Keywords (in text query field) Abstract Text Return: Query Results Return items starting with number Query Form Database: Astronomy Physics arXiv e-prints
This paper analyzes a model of a multiplexer for packetized voice and data. A major part of the analysis is devoted to characterizing the aggregate packet arrival process resulting from the superposition of separate voice streams. This is done via the index of dispersion for intervals (IDI), which describes the cumulative covariance among successive interarrival times. The IDI seems very promising as a measurement tool to characterize complex arrival processes. This paper also describes the delays experienced by voice and data packets in the multiplexer using relatively simple two-parameter approximations.
In a packet voice communication system, packets are fed to a common queue by a number of independent voice sources and are removed from this queue on a first-come-first-serve basis for transmission over a communication link of finite capacity. Each voice source alternates between active periods, during which packets are generated at regular intervals, and inactive periods, during which no packets are generated. In this paper, we discuss three models, a semi-Markov process model, a continuous-time Markov chain model, and a uniform arrival and service model, to assess the queueing behavior of such systems. Numerical results obtained from each of the three models are compared to each other, to results obtained from a discrete event simulation program, and to results obtained from an M/D/1 analysis. Parameters of the model are the average duration of active and inactive periods, the packet generation rate, the communication link capacity, and the total number of voice sources. Conclusions are drawn regarding which models appear to be most appropriate in the parameter ranges investigated.
The performance of an asynchronous transfer mode (ATM) multiplexer
whose input consists of the superposition of a multiplicity of
homogeneous on-off sources modeled by a two-state Markovian process is
studied. The approach is based on the approximation of the actual input
process by means of a suitably chosen two-state Markov modulated Poisson
process (MMPP), as a simple and effective choice for the representation
of superposition arrival streams. To evaluate the cell loss performance,
a new matching procedure that leads to accurate results compared to
simulation is developed. The application limits of the proposed method
are also discussed. The outstanding physical meaning of this procedure
permits a deep insight into the multiplexer performance behavior as the
source parameters and the multiplexer buffer size are varied
This paper describes a new method for evaluating the queue length distribution in an ATM multiplexer assuming the cell arrival process can be assimilated to a variable rate fluid input. The method is based on a result due initially to Bene allowing the analysis of queues with general input. Its extension to fluid input systems is considered here in the case of a superposition of on/off sources. We derive an upper bound on the complementary queue length distribution. The method is most easily applied in the case of Poisson burst arrivals (infinite sources model). In this case, we derive analytic expressions for the tail of the queue length distribution. A corrective factor is deduced to convert the upper bounds to good approximations. Numerical results justify the accuracy of the method and demonstrate the impact of certain traffic characteristics on queue performance.
The performance of a statistical multiplexer whose inputs consist
of a superposition of voice packet streams is studied. The delay for
such a system is analyzed by solving the Σ D <sub>i</sub>/
D /1 queue. The analytic method can be used to find the
approximate mean delay for an arbitrarily large number of trunks and the
approximate delay distribution when the number of trunks is less than
100. An efficient hybrid simulation of the packet voice multiplexer
which can be used to find the delay distribution for a large number of
trunks is presented. In addition, easily computable error bounds for the
present approximation are provided, and the accuracy of the M /
D /1 approximation is investigated
A study is made of statistical multiplexing of voice packets from
a number of packetized voice sources onto a single channel. Each source
alternates between talkspurt (active period) and silence, and packets
are generated during active periods only. The packets are buffered (in a
finite size buffer) when transmission capacity is not available. An
embedded Markov chain model is adopted to analyze the system and a
numerical technique is presented to compute system performance.
Simulation results validate the analysis