TRAFIL

A Tool for Enhancing Simulation TRAce FILes Processing

Christos Bouras

1,2

, Savvas Charalambides

1,2

, Georgios Kioumourtzis

and Kostas Stamos

1,2,3

Computer Technology Institute and Press, N. Kazantzaki Str., University Campus 26504, Rio, Greece

Computer Engineering and Informatics Dept., University of Patras, Patras, Greece

Technological Educational Institute of Patras, Patras, Greece

Center for Security Studies, P. Kanellopoulou 4, 10177, Athens, Greece

Keywords: Simulation Analysis Tool, Trace File, NS-2.

Abstract: NS-2 (Network Simulator) is one of the most popular discrete event simulators used for network simulation.

Trace files produced by NS-2 provide very useful information for post simulation analysis. This paper

presents the architecture and development considerations for a TRAce FILe analysis tool, which intends to

simplify the management of trace files generated during network simulations. The tool is focused on NS-2

trace files, but can be extended to handle a variety of simulation trace files formats. Its purpose is to make

the execution of a large number of network simulations faster, and the extraction of results from a large

amount of data more flexible and productive. TRAFIL introduces a novel way of interpreting, parsing,

reading and eventually utilizing NS-2 trace files by using "metafiles" and "sub-metafiles" during the trace

file recognition and process procedures, making the overall operations more abstract, substantially efficient

and faster than alternative approaches. Furthermore, TRAFIL facilitates the whole trace file analysis task,

offering the opportunity to store each trace file as well as every measurement produced for each trace file.

The tool aims to aid the analysis of simulation results offering features that other tools in this area have been

missing.

1 INTRODUCTION

The simulation of networks (Breslau, 2000;

Fujimoto, 2003) in telecommunications and

computer networks is a technique where the

behaviour of a real network is modelled using

software. Often a simulation of the network’s

functionality is represented by calculating the

interaction of its pieces and then analyzing the

performance.

Network simulation tools (Weingärtner, 2009)

are widely used for simulating physical networks

with great ease. A very popular network simulation

tool is NS-2 (Breslau, 2000; Network Simulator NS-

2; Cicconetti, 2006), which is a discrete event

simulator used for research in computer networks. It

can be used for simulating a variety of protocols

such as multicast, TCP and routing in different kinds

of networks.

Trace files (NS-2 Trace Formats) are considered

some of the most valuable information a network

simulator like NS-2 has to offer, they describe all the

events that have taken place during a simulation and

based on these events users retrieve information

about the network’s behaviour.

There are similar tools to TRAFIL that produce

statistics of a simulated network’s behaviour, but

some of these projects have integrated NS-2, unlike

our tool which is capable to only use NS-2 trace files

to produce the requested measurements (Network

Simulation Cradle Homepage; NS2Measure

Homepage). There are also tools like JTrana(Qian,

2008), Trace Graph(Malek, 2003; Trace Graph

Source Code) and NS-2 Trace Analyzer (Salleh,

2006; Trace Analyzer Source Code) which offer the

opportunity to analyze NS-2 trace files and produce

numerous statistics, metrics and charts.

All of the above tools provide useful information

about a certain simulation scenario at each time and

therefore all the measurements and results refer to

only one trace file. In addition, these tools do not

provide the user with the opportunity to store the

trace file he has analyzed, for instance in a database,

so that he can reuse it without having to reopen it. In

terms of performance when it comes to loading a

Bouras C., Charalambides S., Kioumourztis G. and Stamos K..

TRAFIL - A Tool for Enhancing Simulation TRAce FILes Processing.

DOI: 10.5220/0003976300610064

In Proceedings of the International Conference on Data Communication Networking, e-Business and Optical Communication Systems (DCNET-2012),

pages 61-64

ISBN: 978-989-8565-23-5

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

trace file for analysis the earlier mentioned tools

behave well for small sized trace files. However,

when it comes to serious simulations (Ben-El-

Kezadri, 2008) that produce trace files in the orders

of MB the performance deteriorates significantly.

The motivation for writing this paper is to

present a new post simulation analysis tool that aims

to make the extraction of information from trace

files faster, productive and more flexible. This tool

named TRAFIL gives the opportunity to process as

well as store simulation trace files and then conduct

various calculations and produce measurements

based on their data.

The remainder of the paper is organized as

follows: In section 2 we describe the architecture of

TRAFIL with description of each component

individually. Section 3 presents use cases and results

of TRAFIL usage and finally Section 4 concludes

the paper.

2 TRAFIL ARCHITECTURE

In this section TRAFIL's architecture is presented

along with an explanation of every module that it

consists of. Figure 1 depicts the tool's architecture.

TRAFIL is the main module and it consists of 3

layers, the Presentation layer, the Business Layer

and the Database access layer.

Figure 1: TRAFIL Architecture.

MySQL is used as the local database. Every trace

file is stored in its own table depending on the

metafile and sub-metafiles it was matched to,

because the column names and data type of each

column are extracted from them.

The metafile and sub-metafile repository is the

component of TRAFIL that contains all the metafiles

and their sub-metafiles. They are stored in a

hierarchical order. Each metafile's sub-metafiles are

stored in a unique folder named after the metafile

they belong to. This way the loading of the sub-

metafiles for a specific metafile becomes simpler.

Metafiles are unique files that describe each of

the three main trace formats (NS-2 Trace Formats)

that NS-2 produces which are the Normal trace

format, the Old wireless and the New Wireless.

They are used to encode the structure of a specific

trace file format and to give information regarding

the procedure of generating the metrics for a specific

trace format, namely the names of columns whose

data will be used. Using the metafiles TRAFIL can

match any trace file that belongs to each of the 3

aforementioned categories. Furthermore, metafiles

can be easily extended in order to support any type

of trace file originating either from NS-2 or from

another simulator, thus the use of metafiles renders

TRAFIL a robust tool in regard to coping with

changes of the input’s format e.g. newer versions of

NS-2.

NS-2 trace files can have extra information

depending on the packet types used in the

simulation. This extra information comes in the form

of additional header fields. This way a trace file may

also contain different kinds of information in various

lines. So we could say that each of these 3 main

trace file types also has some sub types depending

on the packets that the simulation uses. Normal trace

files can log header information about TCP packets

and Satellite packets. Old Wireless trace files can

log information about ARP, CBR, TORA, AODV,

DSR, TCP, IP, IEMP and RCA packet types. New

Wireless can log information about ARP, CBR,

TORA, AODV, DSR, TCP, IP and IEMP packets.

The combinations that can appear in a trace file

cannot be known in advance since each simulation is

different from another and the traffic used in each

simulation can be anything the user wants. The

number of header fields in each line is different in

some cases from other lines in the same trace file

and in some other cases so is the order that the

header fields appear. It becomes obvious that the

trace file recognition process is a very complex one.

Each line in a trace file must be carefully examined

to establish what types of header fields are present in

order to handle the information it contains. The

solution our tool uses to tackle this task is the use of

“sub-metafiles”. Sub-metafiles are extensions of the

metafiles that were introduced earlier. For each

metafile we have additional sub-metafiles that

encode the format for every different kind of header

a trace file of a specific format can log. Therefore,

the metafile used to match Normal trace files has

two sub-metafiles, the metafile used to match an Old

wireless trace file has nine sub-metafiles and the

DCNET 2012 - International Conference on Data Communication Networking

metafile used to match a New Wireless trace file has

eight sub-metafiles.

3 TRAFIL USAGE AND RESULTS

3.1 TRAFIL Use Cases

In this section we illustrate an example usage of

TRAFIL where we open a new trace file and present

the information that the tool can extract. The

scenario that was used to create the sample trace file

involved 4 nodes connected in the topology shown

in Figure 2. The simulation lasted 80 seconds and

the communication was between nodes 0-3 and

nodes 1-3. This scenario is a small one and its sole

purpose is to present TRAFIL rather than validate or

infer conclusions about the simulation itself.

Figure 2: Sample Topology.

When the trace file has been analyzed and

processed its contents are visible via the tool from a

table, a feature that is unique in TRAFIL and its

purpose is to enable users to have all the information

they could possibly need centralized and ready to

use. The columns of the table are created based on

the metafile and sub metafiles that were used to

process the trace file. TRAFIL can also produce

specific QoS information about the simulation. The

QoS parameters are Packet Delivery Rate in

packets/second, Throughput in bits/second,

minimum End to End Delay, maximum End to End

Delay, average End to End Delay, Delay Jitter,

minimum Delay Jitter, maximum Delay Jitter,

average Delay Jitter and Packet Loss Ratio. All of

the above QoS parameters are calculated by

selecting two nodes of the simulation and the level at

which the calculation of the parameters will take

place. For a wired scenario the choices are Link

Layer or Network Layer and for Wireless scenarios

the choices depend on the trace levels present on the

trace file and are a subset of AGT, RTR or MAC

layer. The calculation of Throughput is performed

by dividing the number of bits that were successfully

sent and received between the selected nodes at the

time of the arrival of the first and last bit between

the nodes. The End to End Delay for each packet

between the two nodes is calculated by finding the

difference in the arrival time and transmission time

of every packet that was exchanged from the sender

node to the receiver node. The Jitter related

parameters are calculated based on the RFC 3550

(Schulzrinne, 2003) for RTP packets. Finally the

Packet Loss Ratio is calculated by finding all the

packets that were sent and received between the two

nodes, dividing their difference by the packets sent

and multiplying by 100. Finally, TRAFIL enables a

user to create charts regarding Packet Delivery Rate

in packets/second, Throughput in bits/second, Delay

Jitter and End to End Delay for each packet send

between two selected nodes. There are two options

for the time interval in which the sampling of values

takes place in order to create a chart and it can be set

to 1 second or 5 seconds. We believe that these two

options are adequate to create charts with an

acceptable level of detail.

3.2 Results

In this section we present some results regarding the

processing time of TRAFIL when it comes to

opening and storing new trace files of various sizes

to the local database. The results are presented in

Table 1 and they show that, even for very large scale

simulations, TRAFIL manages to minimize the

processing time. It is obvious though that the

increase of processing time is near liner to the trace

file sizes as one would expect.

Another observation extracted from the results is

that for the trace files with sizes 11,413KB and

16,724KB compared to the ones with 25,251KB and

30,774KB sizes respectively, the latter ones have

smaller processing time. This is because the first pair

of trace files have Normal trace format and the

number of sub metafiles used to process them is less

than the number of the ones that would be used if

they had Old Wireless or New Wireless trace

formats. Hence, the execution time, to some extent,

not only depends on the trace file size but also on the

trace format.

Table 1: Processing Time results.

Trace File Size(KB)

Execution Time(ms)

2,524

2516

11,413

16453

16,724

13243

25,251

30751

30,774

21885

42,735

48427

54,462

59974

66,286

73092

115,335

127808

TRAFIL - A Tool for Enhancing Simulation TRAce FILes Processing

4 CONCLUSION-FUTURE WORK

We presented in this work a new tool to enhance

simulations conducted with NS-2. This new tool

named TRAFIL aims at offering an easy to use

interface in which users can load and store

simulation trace files in order to facilitate the

analysis procedure. We used simulation trace files to

produce measurements, statistics and charts in an

effort to help students and researchers when running

simulations with NS-2. TRAFIL was written in the

JAVA programming language and is platform

independent. The main benefit of TRAFIL apart

from the production of ready to use metrics and

charts is the reduced time it consumes for the whole

trace file analysis procedure; firstly by minimizing

the time to open and use a trace file and secondly by

enabling the user to store each trace file in a

database for easy access and future use. That was

something that was missing all previous years in the

NS-2 simulation environment. It also gives the

opportunity to extract all the information that is

stored or produced to a text or Excel file when it

comes to the trace file, simulation metrics and

information or jpeg for the charts. Finally, TRAFIL

can be extended to support a variety of trace files

and not only NS-2 trace files with great ease due to

the use of Metafiles and Sub Metafiles.

In our future work we plan to extent TRAFIL

with useful tools for simulating video transmission

scenarios. The idea is to fully automate the whole

simulation and analysis procedure. In addition to

automating the simulation process we plan to offer

the opportunity to perform SQL queries directly to

the database and returning the results to the user.

This way a user can retrieve all the different

combinations of values contained in a trace file.

TRAFIL is publicly available at (TRAFIL

download).

REFERENCES

Lee Breslau, Deborah Estrin, Kevin Fall, Sally Floyd,

John Heidemann, Ahmed Helmy, Polly Huang, Steven

McCanne, Kannan Varadhan, Ya Xu, Haobo Yu.

2000. “Advances in Network Simulation”, IEEE

Computer,Vol. 33, No. 5, pp. 59 – 67.

R. Fujimoto, K. Perumalla, A. Park, H. Wu, M. Am-mar,

and G. Riley, 2003. Large-scale network simulation:

how big? how fast? Modeling, Analysis and

Simulation of Computer Telecommunications

Systems, 2003. MASCOTS 2003. 11th IEEE/ACM

International Symposium on, pages 116-123, 2003.

E. Weingärtner, H. vom Lehn, and K. Wehrle, 2009. A

performance comparison of recent network simulators.

In Proceedings of the IEEE International Conference

on Communications 2009 (ICC 2009), Dresden,

Germany, IEEE.

Network Simulator NS-2: http://www.isi.edu/nsnam/ns/

Claudio Cicconetti, Enzo Mingozzi, Giovanni Stea, 2006.

An Integrated Framework for Enabling Effective Data

Collection and Statistical Analysis with ns-2. In

Proceedings of WNS2 '06; Proceeding from the 2006

workshop on ns-2: the IP network simulator.

NS-2 Trace Formats: http://nsnam.isi.edu/nsnam/index.

php/NS-2_Trace_Formats

Network Simulation Cradle Homepage: http://www.wand

.net.nz/~stj2/nsc/software.html

NS2Measure Hoepage: http://cng1.iet.unipi.it/wiki/index.

php/Ns2measure

Qian, H. and Fang, W. (2008). Jtrana: A java-based ns2

wireless trace analyzer: http://sites.google.com/site/

ns2trana/

J. Malek and K. Nowak, 2003. Trace graph-data

presentation system for network simulator ns. In

Proceedings of the: Information Systems - Concepts,

Tools and Applications (ISAT 2003), Poland.

Trace Graph Source Code: http://www.angelfire.com/al4/

esorkor/

Trace Analyzer Source Code: http://trace-analyzer.source

forge.net/

Aliff Umair Salleh, Zulkifli Ishak, Norashida Md. Din, Md

Zaini Jamaludin, 2006. Trace Analyzer for NS-2. In

Proceedings of the 4th Student Conference on

Research and Development (SCOReD 2006), Shah

Alam, Selangor, MALAYSIA.

Ryad Ben-El-Kezadri, Farouk Kamoun, Guy Pujolle,

2008. XAV: a fast and flexible tracing framework for

network simulation. Proceedings of the 11th

international symposium on Modeling, analysis and

simulation of wireless and mobile systems, Vancouver,

British Columbia, Canada

H. Schulzrinne, S. Casner, R. Frederick, V. Jacobson,

2003. “RTP: A Transport Protocol for Real-Time

Applications”, RFC 3550, July 2003

TRAFIL download: http://ru6.cti.gr/ru6/research_tools.

php#TRAFL

DCNET 2012 - International Conference on Data Communication Networking