Large-scale network simulation techniques

Yaun, Garrett; Bauer, David W.; Bhutada, Harshad L.; Carothers, Christopher D.; Yüksel, Murat; Kalyanaraman, Shivkumar

doi:10.1145/956993.956999

Cited by 35 publications

(13 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Taking a similar approach for experiment #2 yields an estimate of 439,864 events/second, where the cost of 64-bit processing and larger event lists causes MesoNet to run about 40% slower. These event rates can be compared with those reported (Yaun et al 2003) for RossNet, a parallel simulator for large networks. Simulating synthetic topologies of 4 to 32 nodes on one, two or four instruction streams, RossNet averaged 256,244 events/second.…”

Section: Mills Schwartz and Yuanmentioning

confidence: 99%

“…Several researchers (e.g., Riley et al 2004, Yaun et al 2003, Zeng et al 1998 investigate the use of parallel processing to simulate TCP/IP networks. For example, RossNet (Yaun et al 2003) can simulate large, fast networks with hundreds of thousands of simultaneous flows. Unfortunately, parallel simulation alone does not reduce the parameter space required to explore a wide range of conditions.…”

Section: Related Workmentioning

confidence: 99%

“…Such simulations can be difficult to configure and usually require infeasible resources to explore the parameter space. Several researchers (Riley et al 2004, Yaun et al 2003, Zeng et al 1998) investigate parallel techniques as a means to simulate larger, faster networks. Unfortunately, such techniques do not reduce the parameter space, which remains difficult to configure and continues to require significant resources when conducting careful exploration.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

How to model a TCP/IP network using only 20 parameters

Mills¹,

Schwartz²,

Yuan³

2010

Proceedings of the 2010 Winter Simulation Conference

View full text Add to dashboard Cite

Most simulation models for data communication networks encompass hundreds of parameters that can each take on millions of values. Such models are difficult to understand, parameterize and investigate. This paper explains how to model a modern data communication network concisely, using only 20 parameters. Further, the paper demonstrates how this concise model supports efficient design of simulation experiments. The model has been implemented as a sequential simulation called MesoNet, which uses Simulation Language with Extensibility (SLX). The paper discusses model resource requirements and the performance of SLX. The model and principles delineated in this paper have been used to investigate parameter spaces for large (hundreds of thousands of simultaneously active flows), fast (hundreds of Gigabits/second) simulated networks under a variety of congestion control algorithms. INTRODUCTIONPaxson and Floyd (1997) describe many difficult problems that impede simulation of large data communication networks, and recommend two main coping strategies: search for invariants and carefully explore the parameter space. Unfortunately, typical network simulators (e.g., Fall and Varadhan 2009, SSFNet 2009, Tyan et al. 2009) use hundreds of parameters that can each take on millions of values. Such simulations can be difficult to configure and usually require infeasible resources to explore the parameter space. Several researchers (Riley et al. 2004, Yaun et al. 2003, Zeng et al. 1998) investigate parallel techniques as a means to simulate larger, faster networks. Unfortunately, such techniques do not reduce the parameter space, which remains difficult to configure and continues to require significant resources when conducting careful exploration. In this paper, we describe how to model a modern data communication network, including the transmission control protocol (TCP) and Internet protocol (IP), using only 20 parameters. We implemented the model using SLX (Henriksen 2000) as a sequential simulation, called MesoNet. (Any mention of commercial products within this paper is for information only; it does not imply recommendation or endorsement by NIST.) As we demonstrate, a concise parameter space can be searched efficiently and effectively using sequential simulations deployed in parallel, where each simulation explores a selected configuration of parameters. Elsewhere (Mills et al. 2010), we use MesoNet to study a variety of congestion control algorithms proposed for the Internet. In that study, we perform a sensitivity analysis of the model's parameter space, providing key insights that guide design of the experiments. Here, we discuss only two sample experiments to illustrate the utility and resource requirements of MesoNet.

show abstract

Section: Mills Schwartz and Yuanmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

How to model a TCP/IP network using only 20 parameters

Mills¹,

Schwartz²,

Yuan³

2010

Proceedings of the 2010 Winter Simulation Conference

View full text Add to dashboard Cite

show abstract

“…For example, ns-2 (Breslau et al 2000) (also, the current effort in the ns-3 development) provides a rich collection of network algorithms and protocols. Simulation can also scale up to handle large-scale networks (e.g., (Cowie, Nicol, and Ogielski 1999), (Riley 2003), (Yaun et al 2003), (Bajaj et al 1999)). …”

Section: Network Testbedsmentioning

confidence: 99%

Model-driven network emulation with virtual time machine

Liu¹,

Rangaswami²,

Zhao³

2010

Proceedings of the 2010 Winter Simulation Conference

View full text Add to dashboard Cite

We present VENICE, a project that aims at developing a high-fidelity, high-performance, and highly-controllable experimental platform on commodity computing infrastructure to facilitate innovation in existing and futuristic network systems. VENICE employs a novel model-driven network emulation approach that combines simulation of large-scale network models and virtual-machine-based emulation of real distributed applications. To accurately emulate the target system and meet the computation and communication requirements of its individual elements, VENICE adopts a holistic machine and network virtualization technique, called virtual time machine, in which the time advancement of simulated and emulated components are regulated in complete transparency to the test applications. In this paper, we outline the challenges and solutions to realizing the vision of VENICE. A HARMLESS NECESSARY CATWe classify existing experimental networking research testbeds into physical, emulation, and simulation testbeds. Emulation testbeds combine physical testbeds and emulation, and provide the ability to conduct extensive live experiments. Although emulation allows a certain level of flexibility, the emulated network conditions are inherently restricted by physical limitations (such as nodal processing speed, link bandwidth and latency) and cannot be scaled up beyond the capabilities of the underlying physical infrastructure. Comparatively, network simulation is effective at capturing high-level design issues, answering what-if questions, and investigating complex system behaviors, such as multiscale interactions, self-organizing characteristics, and emergent phenomena. However, simulation fairs poorly in other aspects, particularly in its operational realism. Further, simulation development is both labor-intensive and error-prone; reproducing realistic network topology, traffic and diverse operational conditions in simulation is known to be a substantial undertaking.In this position paper, we describe our ongoing efforts on the VENICE project (Virtual Emulation Network Infrastructure for Commodity Environments), which aims to combine the advantages of simulation and emulation techniques to provide a high-fidelity, high-performance, and highly-controllable experimental platform for large-scale network experiments. The central thrust of this project is a novel model-driven network emulation approach for conducting large-scale network experiments. We pursue holistic machine and network virtualization techniques that allow execution of network simulation models and emulation of real implementations of network protocols and distributed applications under a coordinated virtual time-line. VENICE consists of simulated networks connecting to emulated hosts sandboxed inside virtual machines running real implementations of network applications or network protocols (e.g., router software). Traffic generated from the applications is conducted by the virtual network with appropriate delays and losses according to both simulated and emulat...

show abstract

“…Preliminary success in running massive network simulations on large parallel computers is particularly encouraging [3]. In the same spirit, parallel simulators have been developed as candidates for general network modeling tasks (for example, [1,16,19,20]). …”

Section: Introductionmentioning

confidence: 99%

Parallel Simulation of Hybrid Network Traffic Models

Liu

2007

21st International Workshop on Principles of Advanced and Distributed Simulation (PADS'07)

View full text Add to dashboard Cite

We examine a parallel processing method for simulations of large-scale networks with a hybrid traffic representation combining both a time-stepped fluid model and a discreteevent packet-oriented model. This method benefits from the observation that the time it takes to propagate fluid characteristics along the path taken by the traffic flows has a lower bound equal to the minimum link delay as manifested by the governing ordinary differential equations (ODEs). A better lookahead can thus be used to allow parallel simulation of the hybrid model to run without more synchronization overhead than the corresponding discrete-event packet-oriented model. We derive an analytical model comparing the fluid model and the packet-oriented model both for sequential and parallel simulations. We demonstrate the benefit of the parallel hybrid model through a series of simulation experiments of a large-scale network consisting of over 170,000 hosts and 1.6 million traffic flows on a small parallel cluster.

show abstract

Large-scale network simulation techniques

Cited by 35 publications

References 17 publications

How to model a TCP/IP network using only 20 parameters

How to model a TCP/IP network using only 20 parameters

Model-driven network emulation with virtual time machine

Parallel Simulation of Hybrid Network Traffic Models

Contact Info

Product

Resources

About