Packet Size Distribution Research Articles

Modeling Queueing and Channel Access Delay in Unsaturated IEEE 802.11 Random Access MAC Based Wireless Networks

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> In this paper, we present an analytic model for evaluating the queueing delays and channel access times at nodes in wireless networks using the IEEE 802.11 Distributed Coordination Function (DCF) as the MAC protocol. The model can account for arbitrary arrival patterns, packet size distributions and number of nodes. Our model gives closed form expressions for obtaining the delay and queue length characteristics and models each node as a discrete time <formula formulatype="inline"><tex>$G/G/1$</tex></formula> queue. The service time distribution for the queues is derived by accounting for a number of factors including the channel access delay due to the shared medium, impact of packet collisions, the resulting backoffs as well as the packet size distribution. The model is also extended for ongoing proposals under consideration for 802.11e wherein a number of packets may be transmitted in a burst once the channel is accessed. Our analytical results are verified through extensive simulations. The results of our model can also be used for providing probabilistic quality of service guarantees and determining the number of nodes that can be accommodated while satisfying a given delay constraint. </para>

Distributed sampling measurement method of network traffic in high-speed IPv6 networks

With the advent of large-scale and high-speed IPv6 network technology, an effective multi-point traffic sampling is becoming a necessity. A distributed multi-point traffic sampling method that provides an accurate and efficient solution to measure IPv6 traffic is proposed. The proposed method is to sample IPv6 traffic based on the analysis of bit randomness of each byte in the packet header. It offers a way to consistently select the same subset of packets at each measurement point, which satisfies the requirement of the distributed multi-point measurement. Finally, using real IPv6 traffic traces, the conclusion that the sampled traffic data have a good uniformity that satisfies the requirement of sampling randomness and can correctly reect the packet size distribution of full packet trace is proved.

Performance analysis and comparison of interrupt-handling schemes in gigabit networks

Interrupt processing can be a major bottleneck in the end-to-end performance of Gigabit networks. The performance of Gigabit network end hosts or servers can be severely degraded due to interrupt overhead caused by heavy incoming traffic. In particular, excessive latency and significant degradation in system throughput can be encountered. Also, user applications may livelock as the CPU power gets mostly consumed by interrupt handling and protocol processing. A number of interrupt-handling schemes has been proposed and employed to mitigate the interrupt overhead and improve OS performance. Among the most popular interrupt-handling schemes are normal interruption, polling, interrupt coalescing, and disabling and enabling of interrupts. In previous work, we presented a preliminary analytical study and models of normal interruption and interrupt coalescing. In this article, we extend our analysis and modeling to include polling and the scheme of interrupt disabling and enabling. For polling, we study both pure (or FreeBSD-style) polling and Linux NAPI polling. The performances for all these schemes are compared using both mathematical analysis and discrete–event simulation. The performance is studied in terms of three key performance indicators: throughput, system latency, and the residual CPU bandwidth available for user applications. As opposed to our previous work, we consider not only Poisson traffic, but also bursty traffic with empirical packet size distribution. Our analysis and simulation work gives insight into predicting the system performance and behavior when employing a certain interrupt-handling scheme. It is concluded that no single interrupt-handling scheme outperforms all other schemes under all traffic conditions. Based on obtained results, we propose and discuss a novel hybrid scheme of interrupt disabling–enabling and pure polling in order to attain peak performance under low and heavy traffic loads.

The origin of the skewed amplitude distribution of spontaneous excitatory junction potentials in poorly coupled smooth muscle cells

The skewed amplitude distribution of spontaneous excitatory junction potentials (sEJPs) in the mouse vas deferens and other electrically-coupled smooth muscle syncytia has been attributed to electrically-attenuated depolarizations resulting from the spontaneous release of quantized packets of ATP acting on remote smooth muscle cells (SMCs). However, in the present investigation surface SMCs of the mouse isolated vas deferens were poorly electrically coupled, with input resistances (176±18 MΩ, range: 141–221 MΩ, n=4) similar to those of dissociated cells. Furthermore, the amplitude of evoked EJPs was more variable in surface compared with deeper SMCs (F test, F=17.4, P<0.0001). Using simultaneous electrophysiology and confocal microscopy to investigate these poorly-coupled cells, it is shown that α-latrotoxin-stimulated sEJPs correlate, in timing (median delay ranged from −30 to −57 ms, P<0.05 in all experiments, n=5) and amplitude (Pearson product moment correlation, ρ>0.55 and P<0.001), with purinergic neuroeffector Ca2+ transients (NCTs) in SMCs. The temporal correlation between sEJPs of widely ranging amplitude with NCTs in the impaled SMC demonstrates that all sEJPs could arise from neurotransmitter action on the impaled cell and that the skewed distribution of sEJPs can be explained by the variable effect of packets of ATP on a single SMC. The amplitude correlation of sEJPs and NCTs argues against the attenuation of electrical signal amplitude along the length of a single SMC. The skewed sEJP amplitude distribution arising from neurotransmitter release on single SMCs is consistent with a broad neurotransmitter packet size distribution at sympathetic neuroeffector junctions.

Rate-based pacing for small buffered optical packet-switched networks

Feature Issue on Transmission in Optically Transparent Core NetworksOne of the difficulties of optical packet-switched (OPS) networks is buffering optical packets in the network. O(1) reading operation is not possible in the optical domain, because there is no equivalent optical RAM available for storing packets. Currently, the only available solution that can be used for buffering in the optical domain is using long fiber lines called fiber delay lines (FDL). However, FDLs have important limitations and may cause high packet drop rates due to the burstiness of Internet traffic. We propose an architecture using an explicit congestion control protocol (XCP) based utilization control algorithm designed for OPS wavelength-division-multiplexing (WDM) networks with pacing at the edge nodes for decreasing the buffer requirements at core nodes. We evaluate the FDL requirements on a meshed network with multiple-hop paths and show how FDL requirements change with slot size, utilization, FDL granularity, scheduling, and packet size distribution.

Analytical model for the IPACT dynamic bandwidth allocation algorithm for EPONs

Feature Issue on Passive Optical Network Architectures and TechnologiesWe describe the research that has been performed in the field of the dynamic bandwidth allocation algorithm, interleaved polling with adaptive cycle time (IPACT), for Ethernet passive optical networks (EPON). The main focus has been on modeling packet delay analytically. To derive the packet delay, an important part of the analysis will also focus on the cycle time. It is assumed that the traffic load is symmetric, that packet arrivals are Poisson distributed, and that packets have fixed size. Simulations were performed to prove the accuracy of the analytical model. Some extensions and limitations of the model are treated, including asymmetric traffic load, packet size distribution, self-similar traffic, and differentiated services.

Performance Analysis of IEEE 802.11 DCF in Imperfect Channels

IEEE 802.11 is the most important standard for wireless local area networks (WLANs). In IEEE 802.11, the fundamental medium access control (MAC) scheme is the distributed coordination function (DCF). To understand the performance of WLANs, it is important to analyze IEEE 802.11 DCF. Recently, several analytical models have been proposed to evaluate the performance of DCF under different incoming traffic conditions. However, to the best of the authors' knowledge, there is no accurate model that takes into account both the incoming traffic loads and the effect of imperfect wireless channels, in which unsuccessful packet delivery may occur due to bit transmission errors. In this paper, the authors address this issue and provide an analytical model to evaluate the performance of DCF in imperfect wireless channels. The authors consider the impact of different factors together, including the binary exponential backoff mechanism in DCF, various incoming traffic loads, distribution of incoming packet size, queueing system at the MAC layer, and the imperfect wireless channels, which has never been done before. Extensive simulation and analysis results show that the proposed analytical model can accurately predict the delay and throughput performance of IEEE 802.11 DCF under different channel and traffic conditions

On the effects of the packet size distribution on the packet loss process

Real-time multimedia applications have to use forward error correction (FEC) anderror concealment techniques to cope with losses in today's best-effort Internet. The efficiency of these solutions is known however to depend on the correlation between losses in the media stream. In this paper we investigate how the packet size distribution affects the packet loss process, that is, the distribution of the number of lost packets in a block, the related FEC performance and the average loss run length. We present mathematical models for the loss process of the MMPP+M/D/1/K and the MMPP+M/M/1/K queues; we validate the models via simulations, and compare the results to simulation results with an MPEG-4 coded video trace. We conclude that the deterministic packet size distribution (PSD) not only results in lower stationary loss probability than the exponential one, but also gives a less correlated loss process, both at a particular average link load and at a particular stationary loss probability as seen by the media stream.Our results show that for applications that can only measure the packet loss probability, the effects of the PSD on FEC performance are higher in access networks, where a single multimedia stream might affect the multiplexing behavior. Our results show that the effects of the PSD on FEC performance are higher in access networks, where a single multimedia stream might affect the multiplexing behavior and thus can improve the queuing performance by decreasing the variance of its PSD.

EPON versus APON and GPON: a detailed performance comparison

Feature Issue on Optical Ethernet (OE) Ethernet passive optical network (EPON) efficiency issues in both upstream and downstream directions are discussed in detail, describing each component of the overall transmission overhead as well as quantifying their effect on the system's performance and comparing them with the other existing passive optical network (PON) access systems, namely, asynchronous transfer mode PON (APON) and generic framing PON (GPON). For EPON, two main transmission overhead groups are defined, namely, Ethernet encapsulation overhead and EPON-specific scheduling overhead. Simulations are performed using the source aggregation algorithm (SAA) to verify the Ethernet encapsulation overhead for various synthetic and measured packet size distributions (PSDs). A SAA based an EPON simulator is used to verify both upstream and downstream overall channel efficiencies. The obtained simulation results closely match the theoretical limits estimated based on the IEEE 802.3ah standard. An estimated throughput of 820 to 900 Mbits/s is available in the upstream direction, whereas in the downstream direction effective throughput ranges from 915 to 935 Mbits/s.

Packet-Mode Scheduling with Proportional Fairness for Input-Queued Switches

Proportional fair bandwidth allocation in packet switches is a fundamental issue to provide quality of service (QoS) support in IP networks. In input-queued switches, packet-mode scheduling delivers all the segments of a packet contiguously from the input port to the output port, thus greatly simplifying the design of packet reassembly modules and yielding performance advantage over cell-mode scheduling under certain conditions [1]. One of the important issues of packet-mode scheduling is how to achieve fair bandwidth allocation among flows with different packet sizes. This paper presents an algorithm called packet-mode fair scheduling (pFS) that guarantees each flow a bandwidth proportional to its reservation regardless of the packet size distribution and the system load. Simulations show that our approach achieves good fairness as well as high throughput and low packet delay. Compared to algorithms without fairness mechanism, pFS yields significant performance improvement in terms of average packet delay when the traffic is heterogeneous. A hardware implementation is presented to show that the proposed algorithm has low complexity and the computation can be completed in a single clock cycle, which makes pFS applicable to high-speed switches.

On the effects of the packet size distribution on FEC performance

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.

Distribution of wave packet sizes in microbarom wave trains observed in Alaska

This work reports on a study of the distribution of wave packet sizes contained in intervals of continuous microbarom activity. Microbaroms are a class of atmospheric infrasound that is characterized by narrow-band, nearly sinusoidal, waveforms with periods near 5 s. They are known to be generated by marine storms, presumably through a nonlinear interaction of surface waves, however the detailed analysis of the process is still incomplete. The data analyzed were obtained using the University of Alaska infrasound array of four microphones located in central Alaska. Because of the narrow-band feature of the microbarom signals, the Hilbert transform is applicable as a method for finding phase breaks in the signal. The phase breaks are interpreted as the demarcation of the boundaries of wave packets. When applied to long sequences of microbaroms a broad distribution of packet lengths is found that diminishes monotonically with length and has a mean near 10 cycles and a variance nearly as large. The distribution function decreases exponentially with packet length. The distribution of packet sizes is influenced by the presence of multiple sources and multiple propagation paths between the sources and the sensor array. Identification of individual packets should open the way to a more detailed analysis of microbarom wave trains. After separating the wave train into individual wavelets the intermicrophone correlation is estimated as a function of microphone separation. As has been observed in earlier microbarom studies, a decrease in correlation was observed for microphone pairs orthogonal to the direction of propagation when compared to correlations between microphones spaced along the direction of wave propagation.

Security performance of loaded IEEE 802.11b wireless networks

Existing solutions for wireless LAN networks have been subject to security vulnerabilities and previous study has addressed and evaluated the security performance of IEEE 802.11b wireless networks using single server–client architecture and simple traffic models. This paper describes research into the effect of multiple security mechanisms on the performance of multi-client congested and uncongested networks using a layered security model. It evaluates the interaction between different security layers and their effects on performance (response time and throughput). This research also evaluates the performance effect of different TCP and UDP packet size distributions on secure wireless networks. The benefits of this wireless network study focus on determining ways in which to configure wireless networks such that security requirements can be met in relation to quantifiable performance impact in practical situations.

Sustainable load of fibre delay line buffers

The sustainable load of fibre delay line buffers, defined as the load at which a system with infinite buffering capacity becomes unstable, is discussed. Owing to the particularities of fibre delay line buffers, namely their finite delay granularity, this sustainable load is generally less than 100%. It is further shown that the packet-size distribution has some impact too.

A packet-size aware adaptive routing algorithm for parallel transmission server systems

This paper describes a routing algorithm that assigns each arriving packet to one of several homogeneous parallel servers. The algorithm differs from existing routing policies in that it considers the impact of each arriving packet's size on the average packet delay. It is derived from mimic optimal routing, which mimics the behavior of numerical solutions to a packet assignment problem. It works under the condition that inter-arrival-time and packet-size distributions are unknown, but some possible candidates are known in advance. The algorithm calculates statistics, such as means of these distributions, from input traffic, so it provides not only good performance under stationary input traffic but also adaptive behavior under dynamic input traffic, where the distributions change with time.

Modeling IP traffic: joint characterization of packet arrivals and packet sizes using BMAPs

This paper proposes a traffic model and a parameter fitting procedure that are capable of achieving accurate prediction of the queuing behavior for IP traffic exhibiting long-range dependence. The modeling process is a discrete-time batch Markovian arrival process (dBMAP) that jointly characterizes the packet arrival process and the packet size distribution. In the proposed dBMAP, packet arrivals occur according to a discrete-time Markov modulated Poisson process (dMMPP) and each arrival is characterized by a packet size with a general distribution that may depend on the phase of the dMMPP. The fitting procedure is designed to provide a close match of both the autocovariance and the marginal distribution of the packet arrival process, using a dMMPP; a packet size distribution is fitted individually to each state of the dMMPP. A major feature of the procedure is that the number of states of the fitted dBMAP is not fixed a priori; it is determined as part of the procedure itself. In this way, the procedure allows establishing a compromise between the accuracy of the fitting and the number of parameters, while maintaining a low computational complexity. We apply the inference procedure to several traffic traces exhibiting long-range dependence. Very good results were obtained since the fitted dBMAPs match closely the autocovariance, the marginal distribution and the queuing behavior of the measured traces. Our results also show that ignoring the packet size distribution and its correlation with the packet arrival process can lead to large errors in terms of queuing behavior.

A unified approach to analyze multiple access protocols for buffered finite users

In this paper, a new approach is proposed to analyze the performance of buffered random multiple access protocols called tagged user analysis (TUA). In TUA, each user is modeled as an independent queueing system, which operates at its own equilibrium probability. As a result, the classical queueing theory can be employed to analyze the queueing behavior of each user. On the basis of the behavior of the tagged user, the performance of the whole system is analyzed. TUA is a unified method of analyzing most of random access methods including the heterogeneous random multiple access techniques that are difficult to analyze otherwise. Furthermore, TUA can also handle those systems where user terminals may have arbitrary packet size (in slots) distribution, arbitrary arrival processes such as batch packet arrivals in Bernoulli, arbitrary service principle such as priority queueing, last came first served, and random service. In this paper, in addition to analyzing for the distributions of several performance indices, their mean values are obtained.

Translation of specification units between IP and ATM quality of service declarations

AbstractAn integral part of any network quality of service (QoS) system are its QoS declarations. QoS declarations consist of service classes, parameters, and specification units. QoS declarations are a component of the QoS architecture, as such they are a source of heterogeneity stemming from the fact that different QoS systems may be based on different QoS architectures and thus use different QoS declarations. A particular problem in that domain is the translation of specification units for QoS systems that are based on different forwarding technologies with respect to variable vs fixed packet sizes, i.e. packet vs cell switching. This is a problem that can be dealt with generically such that its solution can be applied to several situations of technically heterogeneous QoS systems like an RSVP/IntServ—or DiffServ—over an ATM‐based system. While straightforward translations have been proposed, we investigate how more efficient translations can be achieved by using a slight but effective modification of existing AAL framing schemes as well as making use of statistical knowledge about packet size distributions. Copyright © 2003 John Wiley &amp; Sons, Ltd.

On a statistical model of network traffic

A nonlinear analysis of traffic measurements involving a layered neural network (NN) (used for reconstruction of underlying process) has shown that the NN trained on a traffic series reproduces the packet size distribution of real data, which fits the lognormal form. The detailed analysis of the traffic measurements has shown that the reason for appearance of this distribution is a simple aggregation of real data. Applying the principal component analysis, the wavelet filtering, Fourier analysis, and statistical criteria we demonstrated that a few leading components form the main part of network traffic, while residual components play the role of small irregular variations that can be interpreted to be a stochastic noise.

Using packet size distributions to identify real-time networked applications

Communication networks, including the Internet, support a wide range of packet-based applications. It is often necessary to know what applications are in use. Traditionally, applications were readily identified by inspection of data held in the packet header such as the destination port number. However, newer and real-time applications cannot always be detected by such a simple investigation and hence other techniques such as packet classification or deep packet analysis have been developed. Deep packet analysis however has significant problems such as its inability to operate on encrypted data packets, and its need to capture specific packets from the traffic stream. The paper considers an alternative approach to the detection of real-time applications. A search was made for a statistical fingerprint derivable from the observable traffic streams generated by such applications. This has been found to be the packet size distribution of the application, and the paper considers this statistic for a range of such applications and network conditions. A detector, based on the described approach, is presented and evaluated using real network traffic.