Project Description

Horizon - Parallel Network Simulator

Discrete event-based simulation is a commonly used evaluation methodology throughout the development process of networked systems. However, it currently faces at least two significant challenges: First, recent advances in wireless communication technology demand highly accurate simulation models, resulting in a steep increase in model complexity and runtime requirements. Second, multi-processor computers constitute the de-facto default hardware platform even for desktop systems, thus providing cheap yet powerful "private computing clusters". As a result, the parallelization of discrete event simulations significantly gained importance and is therefore (again) in the focus of active research.

Model Complexity: Simulation models of wireless networks typically require a considerably more detailed modeling of the lower network layers than models of wired networks. In particular, the wireless channel and the physical layer demand precise models to capture the subtle effects and interactions of advanced wireless communication technologies such as MIMO transmissions or successive interference cancellation. Consequently, simulation run-times increase drastically which in turn hampers the development process and in-depth evaluations.

Parallel Discrete Event Simulation: Being an active field of research for more than two decades, parallel discrete event simulation is supported by a wide range of network simulation frameworks. Despite this tool support, creating a parallel simulation model is still challenging and running simulations on a distributed simulation cluster is complex. At the same time, the increasing number and speed of processing cores in today's commodity hardware makes a higher degree of parallelization very attractive and cost-effective for speeding up network simulation. Nevertheless, a key challenge in parallel simulations, in particular of wireless networks, is the efficient utilization of the available processing power.

In this project we address these challenges by developing a novel parallelization architecture that specifically focuses on the efficient simulation of wireless network simulation models on state-of-the-art multi-core computers. We primarily investigate means of extracting a maximum degree of parallelism from a given simulation model and schemes to achieve a balanced work load across computing cores.

Current State

The foundations of Horizon have been laid in the last years, the basic system is already matured. Today's research focuses on more sophisticated techniques. We increase the scalability of Horizon by combining it with distributed simulation and investigate techniques to increase the knowledge about event dependencies by learning automatically at both run-time and compile-time.

People

Team

Alumni

  • Georg Kunz
  • Olaf Landsiedel
  • Jasper Heuser
  • Alexander Hocks
  • Marc Peiter
  • Thomas Prinz
  • Jan-Erik Rediger
  • Daniel Schemmel
  • Sascha Schmerling
  • Torsten Sehy
  • Simon Tenbusch

Software

Downloads

You can download Horizon from our project website.

Publications

12.
Parallel Expanded Event Simulation of Tightly Coupled Systems
ACM Transactions on Modeling and Computer Simulation (TOMACS), 26(2):12:1--12:26
January 2016
11.
Proceedings of the 18th ACM International Conference on Modeling, Analysis and Simulation of Wireless and Mobile Systems, Cancún, Mexico (MSWiM'15), page 291-300.
Publisher: ACM, New York, NY
November 2015
10.
Proceedings of the 3rd ACM SIGSIM/PADS Conference on Principles of Advanced Discrete Simulation (SIGSIM-PADS’15), London, United Kingdom, page 73-74.
Publisher: ACM, New York, NY
June 2015
9.
Proceedings of the 7th International ICST Conference on Simulation Tools and Techniques (SIMUTools'14), Lisbon, Portugal, page 31-40.
Publisher: ICST, Brussels, Belgium
March 2014
ISBN: 978-1-63190-007-5
8. Praxis der Informationsverarbeitung und Kommunikation (PIK Journal), 35(4):297--304
November 2012
ISSN: 0930-5157
7.
Proceedings of the 26th ACM/IEEE/SCS Workshop on Principles of Advanced and Distributed Simulation (PADS'12), Zhangjiajie, China, page 23--32.
Publisher: IEEE,
July 2012
ISBN: 978-0-7695-4714-5
6.
Georg Kunz, Simon Tenbusch, James Gross, and Klaus Wehrle
5th International Workshop on OMNeT++ (OMNeT++'12), Desenzano del Garda, Italy
Publisher: ICST,
March 2012
5.
Proceedings of the 5th International ICST Conference on Simulation Tools and Techniques (SIMUTools'12), Desenzano del Garda, Italy, page 119-128.
Publisher: ICST, Brussels, Belgium
March 2012
ISBN: 978-1-4503-1510-4
4.
Georg Kunz, Simon Tenbusch, James Gross, and Klaus Wehrle
Proceedings of the 19th Annual Meeting of the IEEE International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS'11), Singapore, page 359 - 368.
Publisher: IEEE Computer Society, Los Alamitos, CA, USA
July 2011
ISBN: 978-1-4577-0468-0
3.
Proceedings of the 4th International Workshop on OMNeT++ (OMNeT++'11), Barcelona, Spain, page 359-366.
Publisher: ICST, Brussels, Belgium
March 2011
ISBN: 978-1-936968-00-8
2.
Proceedings of the 18th Annual Meeting of the IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS'10), Miami, FL, USA, page 172-181.
Publisher: IEEE Computer Society, Los Alamitos, CA, USA
August 2010
ISBN: 978-0-7695-4197-6
1.
Proceedings of the 17th Annual Meeting of the IEEE International Symposium on Modelling, Analysis and Simulation of Computer and Telecommunication Systems (MASCOTS'09), London, UK, page 575-577.
Publisher: IEEE Computer Society, Los Alamitos, CA, USA
September 2009
ISBN: 978-1-4244-4926-2
Export as:
BibTeX, XML
- Impressum | Datenschutz -