Automating Large-Scale Simulation and Data Analysis with OMNeT++: Lession Learned and Future Perspectives
📝 Abstract
Simulation is widely adopted in the study of modern computer networks. In this context, OMNeT++ provides a set of very effective tools that span from the definition of the network, to the automation of simulation execution and quick result representation. However, as network models become more and more complex to cope with the evolution of network systems, the amount of simulation factors, the number of simulated nodes and the size of results grow consequently, leading to simulations with larger scale. In this work, we perform a critical analysis of the tools provided by OMNeT++ in case of such large-scale simulations. We then propose a unified and flexible software architecture to support simulation automation.
💡 Analysis
Simulation is widely adopted in the study of modern computer networks. In this context, OMNeT++ provides a set of very effective tools that span from the definition of the network, to the automation of simulation execution and quick result representation. However, as network models become more and more complex to cope with the evolution of network systems, the amount of simulation factors, the number of simulated nodes and the size of results grow consequently, leading to simulations with larger scale. In this work, we perform a critical analysis of the tools provided by OMNeT++ in case of such large-scale simulations. We then propose a unified and flexible software architecture to support simulation automation.
📄 Content
Automating large-scale simulation and data analysis with OMNeT++: lessons learned and future perspectives
Antonio Virdis, Carlo Vallati, Giovanni Nardini Dipartimento di Ingegneria dell’Informazione, University of Pisa Largo Lucio Lazzarino 1, I-56122, Pisa, Italy a.virdis@iet.unipi.it, carlo.vallati@iet.unipi.it, g.nardini@ing.unipi.it
Abstract—Simulation is widely adopted in the study of modern
computer networks. In this context, OMNeT++ provides a set of
very effective tools that span from the definition of the network, to
the automation of simulation execution and quick result
representation. However, as network models become more and
more complex to cope with the evolution of network systems, the
amount of simulation factors, the number of simulated nodes and
the size of results grow consequently, leading to simulations with
larger scale. In this work, we perform a critical analysis of the tools
provided by OMNeT++ in case of such large-scale simulations. We
then propose a unified and flexible software architecture to
support simulation automation.
Keywords—OMNeT++;
Large-Scale
Simulations;
Data
Analysis; Simulation automation
I. INTRODUCTION
Nowadays simulation is a methodology widely used to drive
the design and to assess performance of different computer
systems. In computer networks in particular, simulation is
widely adopted to drive the design of network or to assess the
performance of existing deployments for provisioning or
troubleshooting. Simulation models are exploited in place of
real measurements or experiments for two main reasons: (i)
simulation models can handle the complexity of such systems,
characterized by many factors or settings that can influence the
performance simultaneously; (ii) they can overcome the
difficulty of studying systems that are distributed over distant
areas and potentially all over the world.
In this context, OMNeT++ has gained popularity as a
mature simulation tool. Especially in the area of networking,
OMNeT++ is widely adopted by scientists and engineers that
can exploit the availability of many simulation models for
different network technologies, both wired and wireless.
Although simulation models are a simplified representation
of actual systems, the increasing complexity of new
communication technologies is currently pushing at a new
different level the complexity of simulation models. Let us
consider as example cellular networks: recent standards, e.g.
LTE and LTE-Advanced, introduced new functionalities to
handle the increasing demand for bandwidth and offer
additional features to end users, with, however, a significant
increase in complexity, which is necessarily reflected in the
simulation models adopted, characterized by an overwhelming
number of parameters, factors and number of simulated nodes.
Simulation models with a large number of factors and
parameters usually imply simulation campaigns with a large
number of different scenarios, aimed at evaluating the impact
of each one on the overall system performance. Even though
some techniques, e.g. factorial analysis [1], might be employed
to reduce the number of scenarios, such simulation campaigns
require a rigorous methodology to execute such large-scale
experiments and, in particular, to analyze properly the large
amount of results produced.
To this aim, software tools are usually employed to support
the researcher to ensure a proper simulation workflow and
eliminate - or minimize - biases or inaccuracies introduced by
human operations, [2]. Specifically, tools that automate the
execution of the simulation workflow and aid the researcher in
the post-simulation analysis are usually employed. In this
context, OMNeT++ already offers several tools and aids:
An effective Graphical User Interface (GUI), which can
be used to automate the execution of simulations. The
end user can plan the simulation campaign through such
interface exploiting an ad-hoc language adopted by
OMNeT++ to configure the simulations and specify their
parameters. The GUI can be used to run the experiments
and monitor their progress through a graphical
representation of network events.
A post-simulation analysis GUI that can be exploited to
visualize data and analyze metrics. Such GUI offers
some basic data analysis operations, which can be
exploited to produce simple graphs from simulation data.
Some command line tools (opp_run) that drive and
automate the execution of simulations without the GUI.
Such tools can be used to run simulations on systems that
lack of a graphical window system, e.g. a cluster or a
server.
A set of tools (scavetool) to export data from simulation
results into different formats suited for external
programs, e.g. Octave, Matlab, etc.
Although such functionalities are offered to automate the
simulation workflow and aid researchers in post-simulation
analysis, they have issues in handling larg
This content is AI-processed based on ArXiv data.