Performance Evaluation of ns-3 Real-Time Emulation
With its real-time mode and tap-bridges, the network simulator ns-3 offers the possibility to integrate real applications or even hardware into a testbed in which ns-3 emulates the network topology. This eliminates the need to develop separate simulation models for the applications. However, there a...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10943112/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850211328863502336 |
|---|---|
| author | Martin Ottens Jorg Deutschmann Kai-Steffen Hielscher Reinhard German |
| author_facet | Martin Ottens Jorg Deutschmann Kai-Steffen Hielscher Reinhard German |
| author_sort | Martin Ottens |
| collection | DOAJ |
| description | With its real-time mode and tap-bridges, the network simulator ns-3 offers the possibility to integrate real applications or even hardware into a testbed in which ns-3 emulates the network topology. This eliminates the need to develop separate simulation models for the applications. However, there are several differences compared to a model-based discrete event simulation, where applications and network topology run together in the simulation process without real-time constraints. When using ns-3 to build an emulation environment, it needs to run in real-time mode, which results in considerable limitations with regard to the achievable throughput and the complexity of the simulated topology. This paper compares the performance that researchers can expect from pure model-based simulations and ns-3-based emulation environments. For this, we evaluate the impact of differently complex network topologies and applications on the simulator throughput in a minimal emulation setup. All measurements are compared with a pure model-based simulation setup. The performance of our setup is measured by the number of possible TCP and UDP flows. With complex network topologies and applications, only a few flows can be handled. In contrast, with simple topologies and applications, the measured throughput of our emulation setup is between 14 Mbps and 120 Mbps. We demonstrate that the achievable throughput can be increased by using optimized ns-3 build profiles and modern CPUs, highlighting, in particular, the impact of packet throughput as a limiting factor for the simulation process. |
| format | Article |
| id | doaj-art-fb722410c1d045a781add4682986f0ea |
| institution | OA Journals |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-fb722410c1d045a781add4682986f0ea2025-08-20T02:09:34ZengIEEEIEEE Access2169-35362025-01-0113595445955910.1109/ACCESS.2025.355547810943112Performance Evaluation of ns-3 Real-Time EmulationMartin Ottens0https://orcid.org/0009-0003-4257-0087Jorg Deutschmann1https://orcid.org/0000-0002-6216-6087Kai-Steffen Hielscher2https://orcid.org/0000-0002-2051-0660Reinhard German3https://orcid.org/0000-0002-9071-4802Computer Networks and Communication Systems, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, GermanyComputer Networks and Communication Systems, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, GermanyComputer Networks and Communication Systems, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, GermanyComputer Networks and Communication Systems, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen, GermanyWith its real-time mode and tap-bridges, the network simulator ns-3 offers the possibility to integrate real applications or even hardware into a testbed in which ns-3 emulates the network topology. This eliminates the need to develop separate simulation models for the applications. However, there are several differences compared to a model-based discrete event simulation, where applications and network topology run together in the simulation process without real-time constraints. When using ns-3 to build an emulation environment, it needs to run in real-time mode, which results in considerable limitations with regard to the achievable throughput and the complexity of the simulated topology. This paper compares the performance that researchers can expect from pure model-based simulations and ns-3-based emulation environments. For this, we evaluate the impact of differently complex network topologies and applications on the simulator throughput in a minimal emulation setup. All measurements are compared with a pure model-based simulation setup. The performance of our setup is measured by the number of possible TCP and UDP flows. With complex network topologies and applications, only a few flows can be handled. In contrast, with simple topologies and applications, the measured throughput of our emulation setup is between 14 Mbps and 120 Mbps. We demonstrate that the achievable throughput can be increased by using optimized ns-3 build profiles and modern CPUs, highlighting, in particular, the impact of packet throughput as a limiting factor for the simulation process.https://ieeexplore.ieee.org/document/10943112/ns-3emulationcomputer networksreal-time performancesimulation |
| spellingShingle | Martin Ottens Jorg Deutschmann Kai-Steffen Hielscher Reinhard German Performance Evaluation of ns-3 Real-Time Emulation IEEE Access ns-3 emulation computer networks real-time performance simulation |
| title | Performance Evaluation of ns-3 Real-Time Emulation |
| title_full | Performance Evaluation of ns-3 Real-Time Emulation |
| title_fullStr | Performance Evaluation of ns-3 Real-Time Emulation |
| title_full_unstemmed | Performance Evaluation of ns-3 Real-Time Emulation |
| title_short | Performance Evaluation of ns-3 Real-Time Emulation |
| title_sort | performance evaluation of ns 3 real time emulation |
| topic | ns-3 emulation computer networks real-time performance simulation |
| url | https://ieeexplore.ieee.org/document/10943112/ |
| work_keys_str_mv | AT martinottens performanceevaluationofns3realtimeemulation AT jorgdeutschmann performanceevaluationofns3realtimeemulation AT kaisteffenhielscher performanceevaluationofns3realtimeemulation AT reinhardgerman performanceevaluationofns3realtimeemulation |