Accuracy of packet loss monitoring over networked CPE

Hasib, Maheen; Schormans, J.A.; Timotijevic, Tijana

doi:10.1049/iet-com:20060271

Cited by 13 publications

(20 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using Markov chains to model the evolution and statistics of the backlog of packets in router buffers is now ubiquitous in network engineering (Bergstrom et al 2002;Vellambi and Fekri 2008;Liu, Sutton, and Collins 2009). Hasib, Schormans, and Timotijevic (2007) developed an analytic solution quantifying the number of probes needed as a function of the level of accuracy required, the characteristics and load of the packet traffic, and the target packet loss probability. This result is similar to that of Lee and Chanson (2002) where a two-state discrete-space Markov chain is used to model a channel in a wireless network.…”

Section: Modeling Packet Communication Network As Markov Chainsmentioning

confidence: 99%

Design of Experiments for Categorical Repeated Measurements in Packet Communication Networks

2011

View full text Add to dashboard Cite

We study the optimal measurement of packet loss and delay in packet networks by treating all measurements as numerical experiments to which we apply the theory of the design of experiments. Specifically we seek to find the optimal times at which to inject survey (probe) packets. Our approach is to model the target node in the packet communication network, an access buffer, as a discrete-time Markov chain. Given that we may only make a limited number of observations, we present a method for optimally designing the observation times for the chain, and derive both exact and continuous optimal designs. Our results show that, for common optimality criteria, measuring at a uniform rate may not be optimal. This has significance for influencing commercial practice as uniform probing is standard. We show how our method may be generalized to Markov systems with larger state space, and describe computational methods to find optimal designs on any system which evolves according to the Markov principle.

show abstract

Section: Modeling Packet Communication Network As Markov Chainsmentioning

confidence: 99%

Design of Experiments for Categorical Repeated Measurements in Packet Communication Networks

2011

View full text Add to dashboard Cite

show abstract

“…At the ACM SIGMETRICS'05, Roughan [17] investigated the inaccuracies caused by correlations between successive probes and derived fundamental limitations. A series of recent papers by Schormans et al, has also explored the potential inaccuracy inherent in packet level measurement, [22], [26], [25], [9], [12], [20], [9]. It has been discovered that even for simple buffering scenarios there are practical load limits beyond which measurement accuracy degrades very rapidly.…”

mentioning

confidence: 99%

“…In [22] queueing analysis was used to show that, for a tail-drop queue, probes would need to be of very similar length to the user packets or else the measured loss probabilities could easily be in error by many orders of magnitude. One of the significant results of the work reported in [9] is that the burstiness of the loss process also has a critical effect, causing any measurement process to require far more samples than would be expected from a naive calculation based only on the target loss probability. In this paper we focus on loss probing as there has already been significantly more research on probing delays (mean delays in particular).…”

mentioning

confidence: 99%

Measurement of packet loss probability by optimal design of packet probing experiments

Parker¹,

Gilmour²,

Schormans³

2009

IET Commun.

Self Cite

View full text Add to dashboard Cite

Packet level measurement is now critical to many aspects of broadband networking, e.g. for guaranteeing Service Level Agreements, facilitating measurement-based admission control algorithms, and performing network tomography. Because it is often impossible to measure all the data passing across a network, the most widely used method of measurement works by injecting probe packets. The probes provide samples of the packet loss and delay, and from these samples the loss and delay performance of the traffic as a whole can be deduced. However measuring performance like this is prone to errors. Recent work has shown that some of these errors are minimized by using a gamma renewal process as the optimal pattern for the time instants at which to inject probes.This leaves the best rate at which to inject probes as the key unsolved problem, and we address this here by using the statistical principles of the Design of Experiments. The experimental design approach allows us to treat packet level measurements as numerical experiments that can be designed optimally. Modelling the overflow of buffers as a 2-state Markov chain, we deduce the system's likelihood function, and from this we develop a technique (using the Fisher information matrix) to determine the upper-bound on the optimal rate of probing. A generalization of this method accounts for the effect of the probed observations interfering with the experiment. Our numerical results focus on VoIP traffic, allowing us to show how this methodology would be used in practice. One application of this is in measurement-based admission control algorithms, where our technique can be used to provide an upper-bound on the rate at which probes should be injected to monitor the loss performance of the target network, prior to making an admit / don't admit decision.

show abstract

“…It has been known that networks repeat the cycle of being congested-empty-congested indefinitely [3,4]. In fact, when we carefully look into the experimental results in the previous section we can find that there exist certain patterns in network latency variations.…”

Section: A Network Latency Groupsmentioning

confidence: 94%

“…Queueing theory is used with a small scale network having the information on individual links involved [4,5]. However, the computation cost grows dramatically as the size of network increases.…”

Section: Introductionmentioning

confidence: 99%

A classification of network traffic status for various scale networks

Park

Kim

2013

The International Conference on Information Networking 2013 (ICOIN)

View full text Add to dashboard Cite

Techniques of network status estimation and traffic prediction are required for network control and user applications in the contexts where a variety of traffic data sources are available. Due to the difficulty of estimating applications' network demands and the difficulty of predicting network load, however, the management of network resources has often been ignored. This paper presents a heuristic of network status classification that has been observed in various scale operational networks. The basic idea of the approach is that network traffic repeat cycles of congested states and that a variation of network latency is strongly correlated with the past history of the latency. We directly monitor network load by continually measuring endto-end network latencies in real operational networks and classify network traffic status with respect to the stability and the burstiness of the latencies. The experimental results showed that the proposed method is capable of evaluating network traffic status and reflecting the related fluctuations.

show abstract

Accuracy of packet loss monitoring over networked CPE

Cited by 13 publications

References 7 publications

Design of Experiments for Categorical Repeated Measurements in Packet Communication Networks

Design of Experiments for Categorical Repeated Measurements in Packet Communication Networks

Measurement of packet loss probability by optimal design of packet probing experiments

A classification of network traffic status for various scale networks

Contact Info

Product

Resources

About