Efficient parametric analysis of performance measures for communication networks (original) (raw)
Related papers
Chapter 9 A Computer and Communication Network Performance Analysis Primer
CHAPTER 9: PRIORITY SCHEDULING I Up to this point we have concentrated on analyzing the mean throughput rate for multiple resource systems, and the mean throughput rate and mean delay bounds for networks of single resource systems. We saw that for many different types of systems typically one single type of resource is limiting the maximum mean throughput rate. The avenues available for improving system performance are to add more resources (perhaps moving the bottleneck resource elsewhere) or scheduling the single resource in an effective manner. In this section we will focus on different techniques for scheduling a single resource, in order to meet delay criteria. In practice a system cannot be exercised continuously at its maximum mean rate of completing work, but rather must complete work at some lower rate in order to meet delay criteria of different types. How much below complete utilization this one resource can operate is the subject of this section (and the realm of queueing theory as a branch of applied mathematics).
Efficient event-driven simulation approaches to analysis of network reliability and performability
International Journal of Modelling and Simulation, 2009
Exactly computing network reliability and performability measures are NP-hard problems, precluding their frequent use in design of large networks. Instead, Monte Carlo simulation has been frequently used by network designers to obtain accurate estimates. This paper focuses on simulation estimation of network reliability and performability. First, a literature survey of existing approaches is given. Then, using a heap data structure, efficient implementation of two previous approaches, dagger sampling and Markov model, are proposed. Two new techniques, geometric sampling and block sampling, are developed to efficiently sample states of a network.
2009
In this paper the MM K k=1 CP P k /GE/c/L G-queue is introduced and proposed as a generalised Markovian node model in telecommunications networks. An exact and computationally efficient solution is obtained for the steady-state probabilities and performance measures. Issues concerning the computational effort are also discussed. The proposed queue is applied to the performance analysis of optical burst switching (OBS) nodes. The numerical results obtained and also the numerical results of a previous model are compared to the simulation results of the OBS obtained using captured traffic traces. We have also introduced negative customers into the model, in an innovative way, in order to account for the loss of packets due to technology limitations of the FDL's (fiber delay loop), which is rather specific to the optical domain. The model is quite promising as a viable performance predictor.
Estimating Message Transmission Time Over Heterogeneous Disrupted Links
2018 IEEE 19th International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), 2018
We consider fragmented message transmission through a heterogeneous chain of several independently disrupted communication links. The message is prepared for fragmentation before transmission by dividing it into blocks of constant size. In this setting, we derive an approximation for the mean and standard deviation of fragmented message transmission time when one of the links in the heterogeneous chain is much slower than the rest.
Stochastic Models for Throughput Analysis of Randomly Arriving Elastic Flows in the Internet
2002
This paper is about analytical models for calculating the average bandwidth shares obtained by TCP controlled finite file transfers that arrive randomly and share a single (bottleneck) link. Owing to the complex nature of the TCP congestion control algorithm, a single model does not work well for all combinations of network parameters (i.e., mean file size, link capacity, and propagation delay). We propose two models, develop their analyses, and identify the regions of their applicability.
An analytical model for loss estimation in network traffic analysis systems
Journal of Computer and System Sciences, 2006
Simulation models have been developed in order to foresee characteristics of networks, systems or protocols when carrying out tests in laboratories is very expensive or even impossible. This paper presents a simulation model of a multiprocessor network traffic analysis system. The model, which is based on closed networks of queues, evaluates the efficiency of the system depending on the hardware/software platform features. Therefore, this model is able to estimate performance early in the design and development stages simulating a multiprocessor architecture in charge of analysing network traffic. The goodness of the model will be checked by comparing analytical results with practical ones obtained in laboratory using a traffic analysis system that runs on a multiprocessor platform.
Increasing the accuracy of network simulation experiments
2000
Statistical techniques are extremely useful in system performance evaluation. Since any statistic cannot be guaranteed to give a close estimate for every sample, we must design statistics that will give good results on the average or in the long run. Despite of even higher performance computers, quantitative steady-state simulation of even a moderate complex system takes very long time, since to obtain reasonably stable results, that is, estimates with reasonably variances, very larges samples are usually necessary.