Simulating Stochastic Processes with OMNeT++

Abstract

Modeling of correlated traffic processes is a challenge in simulation modeling but is necessary to obtain realistic simulation models of many computer or communication networks. We present a new simple module denoted as ArrivalProcess to be included as a traffic source in OMNeT++ simulation models. The new traffic source supports different stochastic process models which are defined using a XML format. The process descriptions can be automatically generated from trace data using the software tool ProFiDo which is also briefly presented. By means of examples it is shown that the use of stochastic processes rather than simple distributions results in much more accurate results for systems with correlated arrival streams.

Simulation of Adversarial Scenarios in OMNeT++ - Putting Adversarial Queueing Theory from Its Head to Feet

“…Also, a distinguishing feature of this work is that we validate our implementation by a comparison to previous analytical results and make the complete source code available to the public. We facilitate the implementation of new adversaries and reuse of existing ones by a modular codeexcerpt structure and are able to visualize network behavior through the capabilities of OMNeT++ Another related approach is concerned with more complex inter-arrival distributions for traffic generation [16]. By implementing several stochastic process types as modules for OMNeT++, the authors are able to simulate positive correlation between packet arrivals.…”

Traffic modeling with a combination of phase-type distributions and ARMA processes

“…In Kriege and Buchholz (2011b) a module called OMNeT++ Arrival Process Module is presented that generates random numbers from Markovian Arrival Processes, ARTA processes and ARMA processes and has been subsequently extended to generate random numbers from CAPPs. The generation of random numbers from ARMA models and the initialization step described above are already supported by the module from Kriege and Buchholz (2011b) and therefore it only has to be extended by a class for CAPPs that implements the transformation U t = Φ(Z t ) and the generation of the next Y t from U t (cf. Fig.…”

Matching marginal moments and lag autocorrelations with MAPs

“…They can be efficiently applied in several fields of performance evaluation of various systems. Once the traffic is represented by a MAP, it is easy to generate random samples from it ( [11], [13]), which is beneficial in simulations and measurement based performance analysis. It is also well known that queueing models where the arrival and/or service process are given by MAPs can be analyzed in a numerically efficient way by matrix geometric methods [14].…”