SN Computer Science (2023) 4:269

https://doi.org/10.1007/s42979-023-01684-6

ORIGINAL RESEARCH

Tools and Techniques for Teaching and Research in Network Design

and Simulation
Muhammad Azizur Rahman1 · Algirdas Pakstas2

Received: 20 October 2022 / Accepted: 10 January 2023 / Published online: 17 March 2023
© The Author(s) 2023

Abstract
Computer networking is taught for some time at universities. A network is a complex mix of applications, communications
protocols and link technologies, traffic flows and routing algorithms. Teaching the networking concept, the network design
process is a challenging task, requiring designers to balance user performance expectations with costs and capacities. One
of the obvious approaches employed to deal with its complexity is through the use of modelling and simulation techniques.
This paper examines three best educational tools that can be used in the Computer Networking course at Cardiff Metropoli-
tan University or any University for educational purpose. The paper demonstrates that a successful teaching programme
needs to combine a range of teaching tools to achieve its aims. Analysing various aspects of the course provides evaluation
of the overall success of the tools. Many network design and simulation tools have been studied. Based on our research and
experiment, the best educational tools were selected that could be used for the purpose of teaching in the area of computer
communications networks. The wide area network modelling tool Delite, the network simulator ns-3, the topology generation
tool Brite can be used together to teach students the networking concepts. These tools can not only help students to under-
stand computer network principles but also improve their network practical skills. Currently there are various tools available
that are created by different companies and groups of researchers in academia and that are intended for use as practical and/
or educational tools for network design including security issues. To make this task easy, different users, researchers and
companies have developed different network modelling and simulation (MS) tools. Very often, such tools are addressing
different issues in the network design process which are complementary to each other. It would be useful for learning and
improving the network design process by allowing the use of a variety of tools.

Keywords Network simulation · Design · Tools · NS · Delite · Brite

Introduction

Teaching computer network is very challenging because

of the abstract concepts. In the teaching of basic and
advanced computer networks concepts (routing, mobility,
This article is part of the topical collection “Cyber Security and load balancing, reliability, capacity, delay and so on), stu-
Privacy in Communication Networks” guest edited by Rajiv Misra, dents not only have to acquire the theoretical knowledge
R K Shyamsunder, Alexiei Dingli, Natalie Denk, Omer Rana, but also the practical skills to configure and manage real
Alexander Pfeiffer, Ashok Patel and Nishtha Kesswani.
computer network scenarios. Successful course delivery
* Muhammad Azizur Rahman is a key aim for all educators. The failure of only one part
mrahman@cardiffmet.ac.uk of a course can have disastrous effects on the attitudes of
Algirdas Pakstas students. The most of the Universities have been deliver-
algirdas.pakstas@gmail.com ing computer networking course that aims to teach stu-
dents the basics of computer networking. Our experience
1
Cardiff School of Technologies, Cardiff Metropolitan has been that to provide students with a positive learning
University, Western Avenue, Cardiff CF5 2YB, UK
environment requires the development and integration of
2
Communications Software Engineering, Vilnius University, a large number of teaching tools. No one tool can be a
Universiteto g. 3, 01513 Vilnius, Lithuania

SN Computer Science
Vol.:(0123456789)
269 Page 2 of 15 SN Computer Science (2023) 4:269

perfect for teaching in the same way that no one tool will To understand the details of the networking issues, stu-
ever be the perfect for computer networking. The teaching dents can use the Delite for the calculation of the Capacity,
has culminated in a strong emphasis on the student being reliability and delays of the networking module. On the other
an active learner, as opposed to a passive receiver of infor- hand, to see the dynamic behaviours, NS-3 can be used.
mation, which is then regurgitated back to the assessor. Finally to know the network topology, Brite tool case be
The network design process is a challenging task, used. We found one article regarding the tools for teaching
requiring designers to balance user performance expecta- network [7].
tions with costs and capacities [1]. External factors, such The purpose of this paper is to show how simulation tools
as government policies and regulations, the competitive and technology can serve as an effective educational tool
situation, available technological services and products to enhance the learning of process of networking concepts.
are adding complexity to the design process. Organisa- In this paper we examine the three teaching tools that can
tional strategies, culture and policies also affect the plan- be used in the course of computer networking, features and
ning and design process. The amount of human and tech- design process in depth, pedagogical tool for the teaching of
nical resources in the data communication functions of computer network design. If a student needs to understand
the organisation can also strongly affect the choice. Elec- the basic magic works, then they can use Delite; whereas
tronic communication is so ubiquitous in modern business when student needs to see dynamic characteristic of net-
that it is hard to develop an overall strategic vision that work model, then NS-3 would be the suitable for teaching.
is comprehensive and at the same time detailed enough To work and gain knowledge about the topology, the Brite
to be useful. Additionally, these factors can change and would be the perfect tool.
make the network learning and planning process even
more complex. The business role of the proposed network
application adds extra complexity to planning process. Problems in Teaching Computer Networks
To improve the quality of undergraduate education such
a complex topics network simulation tools best option. The Open System Interconnection model (OSI) and com-
There are many authors who attempted to solve the ques- munications network simulation are two parts of the impor-
tion on how can students’ motivation and interest for the tant area of computer networking course. The OSI model
complex teaching material of Computer Networking and provides an abstract function decomposition of a network
other complex module(s) be improved. Currently, many protocol and is commonly used to depict network functional-
works deal to improve the pedagogy of computer network- ity of each layer. There are many tools for network analysis
ing concepts using network design and simulation tools including cost, reliability, capacity, delay etc. both in com-
and techniques [2–6]. The article [7] shows how to reduce mercial and educational sectors. It is easy for the students to
the gap between the education in universities and real work build the visual connection between OSI model and network
environments in the area of computer networking using protocols of each layer.
Packet tracer, GNS2, OPNET, and fat-tree simulator tools. The technical complexities of computer networks and
There are some recent published papers dealing with the data transmission processes and protocols present immense
complex subjects in the classroom [8–10]. challenges for teaching and learning in computer network
Capacity, Reliability, and Delay are three "magic words" education [12]. Students often need to have intensive stud-
used for describing concepts of the computer network design ying and practical experience with hands-on activities to
[11]. While capacity and delay are relatively well under- understand the large numbers of complex and abstract con-
stood, this is not the case for reliability. The term reliability ceptual details on network payers, protocols, addressing
is often used interchangeable with availability and dependa- format and configurations [13]. The traditional mainstream
bility. Increased redundancy in the network structure is often teaching method of lecturing is found to be inadequate for
considered as important, or even the only way to achieve motivating students in network education [4]. The com-
higher levels of reliability. Factors affecting performability mon challenge often seen in teaching computer network-
of the communications systems are discussed. Reliability in ing and data communication is the difficulty for students to
the network design process in a context of systems approach understand the complex technical concepts and to visualise
to network design is considered in more detail. Examples the various network protocols, roles, and functions in data
of calculations of reliability figures for different network communication [15]. More importantly, students may have
designs are considered using the Delite tool. Reliability sup- further difficulty in developing the skills and ability to use
port in communications protocols is discussed. Example of higher level skills for troubleshooting and creative problem
simulations and animations of network dynamics is pre- solving [16].
sented using the ns-3 and nam tools. Example of topology Some issues of the computer networking could be
handling is Brite tool. found in [17]. Computer Networking is difficult to learn.

SN Computer Science
SN Computer Science (2023) 4:269 Page 3 of 15 269

It advances very rapidly. Many Computer Networking • Typically provides an interactive graphical user interface
topics have also quickly developed into separate subjects, for network editing, capacity calculation, traffic estimate,
notably Wireless and Mobile Networks, Optical Networks, performing network optimisation, allowing users to per-
and Network Security. It becomes increasingly difficult to form “what-if” analysis and many other features;
teach the subject effectively, partly because the Internet’s • Displays the results through menus, tables and graphical
success has attracted students with various backgrounds. displays via animated utilisation bars, graphical statistics,
Besides the issues, there are quite a number of obstacles diagrams, reports, charts and graphs.
to teaching and learning Computer Networking effectively, • Analyses the network model produced by designer;
as outlined below. (a) The principles underlying Computer • Recommends the configurations of existing switches (e.g.
Networking are intrinsically very complex. (b) Computer how many access cards to add) and the optimal locations
Networking concepts and protocols are also very abstract of new switches/routers;
to many students. (c) Unlike Computer Programming and • Shows which links and devices are over and underused;
Computer Architecture courses, for example, resource pro- • Predicts link and device throughput and utilisation;
vision for hands-on practical experience in Computer Net- • Shows effects of a link or device failure;
working is problematic. A Computer Networking laboratory, • Shows how changes in topology, device upgrades or
if available, usually has a ready-to-use computer network on expansion will affect performance.
which students can conduct various experiments. However,
students should also be expected to know how to set up a (b) Considering cost-related issues:
computer network from scratch after taking the course, very
much like knowing how to write programmes after taking a • Analyses the status of the current network, evaluates the
Computer Programming class. But it is not quite possible to condition of any user-defined configuration and optimises
provide adequate resources to achieving that in many aca- the expansion of the network based on demand and costs;
demic environments. (d) Many terminologies and acronyms • Ensures that the network is cost effective, reliable and has
are introduced and used in the field, and some of them are sufficient capacity to meet current and expected customer
very similar, e.g. ARP and ARQ. (e) Some of the networking demand.
problems are difficult to comprehend and appreciate due to
students’ common lack of practical experience. (v) Investigation of the network performance by model-
ling and simulation of dynamic traffic behaviour:

• Finds out the limitations of the network as designed by

Students and Researchers Expectations increasing traffic load and investigating its performance;
from the Network Modelling Tools • Identifies the applications or protocols contributing the
most to the network congestion and delay;
Use of network modelling and simulation tools is an impor- • Predicts the response times of applications before deploy-
tant element in the process of design, deployment, testing ing the network;
and evaluation of any communications network. Tools are • Predicts how many users an application can handle;
the integral parts of the overall methodology of planning • Confirms whether the service level agreements can or
and designing a network. Designers may analyse the output cannot be met.
produced by the tools, compare several designs and select
the best design according to some predetermined criteria. Every tool uses its unique native notation to represent
Tools are also needed to study and test new protocols. a network model. The users of the network design tools
Typically, every network design undergoes some standard are network managers and engineers, network designers,
stages and it is expected that tools may help in these stages. researchers, and the students [10–19]. The following are
Network modelling and simulation tools may help [18] in: types of network modelling tools available [20, 21].
(a) The planning of the network topology and link (a) Main design tools:
capacities:
• Analytical tools;
• It is expected that the tool will use only the links with a • Simulation tools.
standard range of capacities.
• Means in the case of network design and upgrade, a tool (b) Supplementary tools:
can interface with existing database systems to obtain the
current network configuration directly without data entry • Topology generation tools;
by designer; • Topology discovery tools.

SN Computer Science
269 Page 4 of 15 SN Computer Science (2023) 4:269

Three of the educational tools are Delite (analytical tool), Features of the Delite, NS‑3 and Brite
NS-3(simulation tools) and Brite (topology generation tool).
Students can use the Delite for the calculation of the capac- Delite
ity, reliability and delays of the networking module. On the
other hand, to see the dynamic behaviours, NS-3 can be used. Delite [22] is an educational and practical wide area net-
Finally to know the network topology, BRITE tool case be work (WAN) design tool, which can produce network
used. designs of limited size using a set of embedded network
Students can get clear concepts using topology generation design algorithms. Data for such designs have to be read
tool(s). They will experiment different topology algorithms from the specially formatted external files. Delite can pro-
using Brite tool. The students can design the topology using duce graphical displays representing network nodes and
menu bar in Brite and it will generate a text file. Students can links as well as some additional analysis data (delay analy-
study the text file and change the parameters. They can design sis and reliability analysis) in special windows. An input
the same model (topology) in Delite tool and perform capac- file contains data for the network design problem: node
ity, reliability and delay analysis using predefine algorithms. coordinates, link parameters, traffic parameters, as well as
Delite also produces the text file and students can edit the file cost and equipment specifications.
and parameters and perform experiments. And finally they Delite provides the network planner with a user-friendly
can re-map the same model in NS-3/Ns-2 and check different menu-driven interface. It uses a 1000 by 1000 unit size
paraments and dynamic behaviour and conduct performance window for a square world on the screen for displaying
analysis. For example, in networking course, a full course- the network model and a separate special type of window
work/assessment can be completed using 2 and/or 3 of these called the ‘Delite Monitor’ for user responses and analysed
tools together based on the requirement of the assessment and output of the designed network model. The origin of the
the module. window is in the top right corner but the distance from
From the assessment point of view, the sequence is: it is measured down and to the left, i.e. the opposite way
Topology Generation → Performance Analysis → Dynamic Behaviour Analysis.

From the tool point of view, the sequence is: from normal X, Y coordinates. An alternative is to put the

Assessment Part 1 (in Brite) → Assessment Part 2 (in Delite) → Assessment Part 1 (in NS − 3).

These tools will also be very useful for the network designer
origin in the bottom left corner and make all values nega-
and researcher in their network design process or algorithm
tive from this origin. Delite supports different types of
development. The use of network modelling and simulation
communication links, which are T1, D96, and D56.
tools is an important element in the process of design, deploy-
For each network model, Delite produces seven text
ment, testing and evaluation of any communications network.
files:
Tools are the integral parts of the overall methodology of
planning and designing a network. Designers may analyse the
• .INP file contains information that refer to the other
output produced by the tools, compare several designs and
files used for generating the network file;
select the best design according to some predetermined cri-
• .GEN file contains sites/nodes, their coordinates, type
teria. Tools are also needed to study and test new protocols.
of link and some parameters;
In the following sections, we will discuss details (related
• .REQ file contains traffic information (SOURCE, DEST
files and their format, steps to design network model, etc.) of
and BANDWIDTH);
these 3 tools which are Delite, Ns-3 and Brite. These can be
• .CST file contains tariff information (end to end cost);
introduced to the students that might be useful for the students
• PARMS.TBL contains several network parameters (e.g.
to have basic networking concepts.
utilisation, APHA, MESEEEHE_LEN, DUPLIXITY,
etc.);

SN Computer Science
SN Computer Science (2023) 4:269 Page 5 of 15 269

• EQUIP.TBL/NOEQUIP.TBL contains equipment infor- designs may have different costs, delays, reliability, aver-
mation (fixed), specifies the cost and capacity of the age number of hops, etc. Other files are used for different
equipment that can be placed at every site; purposes: noeqip.tbl and param.tbl are predefined (used to
• .NET file contains the network topology (e.g. source, supply some network parameters), the.cst file describes the
sink, node coordinates, etc.). This file receives the nec- costs associated with link types (e.g. T1, T3, D96, etc.), the.
essary information from the.INP file. req file describes the capacity of each link type, and the.
inp file contains the names of files related to a particular
The WAN design tool Delite produces a performance model. The users can edit all of the files mentioned above
report of the network model. The report includes average as they are in ASCII text format. The link direction of the
delays of the link, the average number of hops, link utilisa- arrow between file types means the direction of a “call” for
tion for every separate link between nodes, utilisation of information, e.g. the.net file is only using.inp files and the.
each node, overall network model utilisation, etc. A separate inp file is using all others.
special window named a ‘Legend’ is used for showing the The file processing steps are shown in Fig. 2. The.gen file
output. is designed and/or edited by the user and fed to the Delite
The algorithms used for designing the topology of a net- tool. As a consequence,.inp,.cst, and.req files are produced
work model supported by Delite are: using some parameters from para.tbl or noeqip.tbl files.
Finally using these files (.gen,.inp,.cst,.req), the design file
• The Prim Algorithm; (.net) is produced. As the.net file contains the whole network
• The Primdijkstra Algorithm; model information, this file is used by the NeDaSE tool for
• The Tour Building (The Nearest Neighbour) Algorithm; transformation purposes from Delite to other formats. The.
• The Tour Building (The Farthest Neighbour) Algorithm; cst and.req files are not used for transformation (from Delite
• The Esau–Williams Algorithm; to other tools’ formats). This is due to the fact that there is
• The Sharma Algorithm; no way to use the data of.cst (all possible combinations of
• The Multispeed Tree Design Algorithm; link costs) or.req (all possible combinations of link band-
• The Nearest Neighbour Esau–Williams Algorithm; widths) files in other tools. All of the files are in text format
• The Multicenter Esau–William Algorithm; and students can easily change/amend and learn about the
• The Mentor Algorithm; parameter and therefore networking issues.
• The Incrementour Algorithm. The.net file carries the model information. There are three
sections in the.net file [22]. The first section (Fig. 3) contains
Files and Model Processing Steps in Delite node/site information, coordinate of the nodes with some
predefined parameters etc., the second section (Fig. 4) con-
There are seven text files in Delite as shown in Fig. 1. The tains the topology information and predefined parameters
figure also shows the relationship among the files. The most [22] and the third section contains parameter information
important for the users are files.gen (original node informa- (not shown). Furthermore, other files (.cst,.gen,.inp,.req) are
tion, coordinates and available link types) and.net (model also needed to make a model workable in Delite. For this
node coordinates and additionally a table of links between size of thesis, the structures of these files are not provided.
nodes, i.e. actual network design). Links between nodes can Among these files,.cst and.req are two of the biggest files.
be generated using a few design algorithms. Thus, various

Fig. 1  Relationship of model

files (7 files in total) used in
Delite tool

SN Computer Science
269 Page 6 of 15 SN Computer Science (2023) 4:269

Fig. 2  Model processing steps

in Delite tool staring with step 1
to step 4 in sequence

Fig. 3  An example of the node information from Prim7UK model in Delite tool showing the nodes details (standard format)

Fig. 4  An example of the link information from Prim7UK model in Delite tool showing the link details (standard format)

In.gen file (not shown), there are three sections: node the site) and traffic out (leaving the site) of each node. The
information, used link type, and parameter section. The node second section contains link types that are to be used and
section carries the node name, initial node position (before their associated information, and the third section contains
applying any algorithm for a model), parent node, popula- several parameters.
tion concerned with a node (number of users), node level The traffic file.req contains the capacity of each of the
(the level of the site in the hierarchy), traffic in (entering possible links. These possible links are basically the links
(combination) from each node to rest of the nodes of the
model. If there are n nodes, then the total number of links

SN Computer Science
SN Computer Science (2023) 4:269 Page 7 of 15 269

with a bandwidth of the.req is ((n – 1) + (n – 2) + … + (n– Link

n)). The links of a model are duplex.
The file.cst contains the costs of all of the links available Traffic
for the design. If there are n nodes then there will be n + ( Does not allow for specifying the use of the communica-
n – 1) + (n – 2) + ··· + (n–n) links in this.cst file. tion protocols.
The file para.tbl (predefined and used as a template) con- Visualisation
tains information to guide the design process. The visualised location of the node is always the same.
The file noeqip.tbl/eqip.tbl (also predefined and used as Others
a template) describes equipment, e.g. multiplexers, rout-
ers including the cost and capacity of the equipment that 1. No animation is supported.
can be placed at every site [14], p403]. This specifies the 2. In the.net/.gen file, there are a lot of parameters that are
throughput of the equipment and the number of lines that used for the calculation of different factors (e.g. delay
can be terminated. The throughput for data switches is usu- analysis, traffic analysis, reliability analysis, etc.).
ally measured in packets/seconds, and the voice switches are 3. Has to be fitted within 1000 by 1000 “square world”.
measured in call setups/seconds. 4. Performance analysis is done by this tool.
Finally, the.inp file contains the names of all of these files. 5. The models designed by the users are confined within
the predefined algorithms allowed by the tool.
Network Model in Delite
Figure 5 shows the visualisation screen of the Delite tool.
Features of the network Delite model are: The left lower side is the minimum points (0,0) and right
Nodes upper corner is the maximum point/coordinate (1000, 1000)
of the display.
1. By default, the nodes are labelled with N1, N2, N3 and
so on. The first node starts labelling with 1 rather than NS‑3/NS‑2
0.
2. The location of the node is represented by x, y coordi- NS-3 [23, 24] network simulator, a VINT (Virtual Inter-
nates. Network Testbed) project from U.C. Berkeley/LBL/Xerox
3. It supports text format for a node name that may have PARC, is a discrete event simulator targeted at network
more than one word (e.g. United States). research, which provides substantial support for the simula-
4. The minimum coordinates of the square world (visu- tion of TCP, routing and multicast protocols. The simulator
alisation) of the Delite are in the left-down corner and is written in C +  + and uses a Tcl/Otcl as a command and
the maximum are in the right top corner of the window/ configuration interface. The ns-3 has the advantage that it
screen (shown in Fig. 5). provides numerous frameworks as the basis for extending
5. The number of nodes is limited to 100. its core capabilities. Ns-3 is derived from REAL (Realistic
6. The.net model is designed from the.gen file. During the and Large), which is derived from NEST (Network Simula-
editing of the.net file, the node locations can be changed tor testbed).
by the designer. Thus, the coordinates of a node may be In ns-3, arbitrary network topologies, composed of routes,
different in.gen and.net files. Thus, in the case of trans- links and shared media can be defined. A rich set of proto-
forming other models (e.g. ns-3) to the Delite model, it cols (e.g. TCP, TCP/Reno, TCP/Newreno, TCP/Sack1, TCP/
is necessary to generate both.gen and.net files. Fack, TCP/FullTcp, TCP/Vegas, TCP/Vegas/RBP, TCP/
Vegas/RBP, TCP/Asym, TCP/Reno/Asym, TCP/Newreno/
Asym, TCPSink, TCPSink/DelAck, TCPSink/Asym, TCP-
Sink/Sack1, TCPSink/Sack1/DelAck, UDP, RTP, RTCP,
LossMonitor, IVS/Source, IVS/Receiver, CtrMcast/Encap,
CtrMcast/Decap, Message, Message/Prune, SRM, SRM/
Adaptive, Tap, Null, RtProto/DV) are available and various
types of applications can be simulated. Among them are
FTP, Telnet, and HTTP, which use TCP as the underlying
transport protocol and applications requiring a constant bit
rate (CBR) traffic pattern, Ih use the UDP transport protocol.
Multiple queuing and scheduling policies can be configured
Fig. 5  Coordinate system of the Delite Model (GUI) showing the such as drop tail (DropTail), random early detect (RED),
maximum and minimum points explicitly congestion notification (ECN), priority and fair

SN Computer Science
269 Page 8 of 15 SN Computer Science (2023) 4:269

Fig. 7  Relationship of model files used in ns-3 tool that shows the
linkage between the input/output files

Files and Model Processing Steps in ns‑3

There are four files (see in Fig. 7) in ns-3 whose processing

Fig. 6  The ns-3 Network Model steps in the tool are shown in Fig. 8. The.ns and.tcl are input
files, and.nam and.tr (trace file) are the output files. The.ns
queuing. The simulator is event-driven and runs in a non-real or.tcl file is given to the ns-3 tool as an input and the.nam
time fashion. Packet losses are simulated by buffer overflows (and/or.tr) file is produced. The.tr file can be used for post-
in routers, which is also the dominant way packets get lost processing. This nam file is animated by the nam anima-
in the Internet. There is also support for error models other tor which again can produce a.ns file. Initially, the student
than losses through buffer overflow. The routing model can will prepare a.ns/.tcl script for the network model using tool
be static or dynamic. The link state and distance vector rout- command language. If students are unfamiliar, then s/he can
ing algorithms are supported. use GUI in nam animator and design the network model
The network topology is specified by describing the net- (drag and drop). In this case, the Nam animator will generate
work nodes and edges [25]. The network model represents a.ns file automatically and the student can open and see the
the interconnection of network elements, hereafter called code(s) of the file (text format) and learn. Then the model
nodes. It consists of nodes and links shown in Fig. 6. can be execute by the nam which will execute the codes and
In ns-3 model, single or multiple traffic generators, simulate the model.
including statistical generators and other typical generators, The main file format that works behind different kinds of
such as FTP and Telnet, can be attached to any node. In addi- network topologies in the ns-3 tool is shown in this section.
tion, network and transport protocol behaviours are simu- The ns-3 file (.ns) normally has six types of information
lated by attaching the appropriate agents to the interested which are; node information (Fig. 9), topology information
nodes. Examples of routing agents include the ones sup- (Fig. 10), protocols (agent) (Fig. 11), data source (applica-
porting static routing (Dijkstra’s SPF algorithm), dynamic tion) (Fig. 12), connection (agent to agent) (Fig. 13) and
routing (Distance Vector Algorithm) or multicast routing. simulation timing information (Fig. 14). Details can be
Examples of the transport agent include the TCP agent and found in this reference[15]. Users can write the code using
the UDP agent. Links are modelled on either the simplex TCL (Tool Command Language) scripting language to
link or duplex link with a predefined capacity, delay and design the network model and execute the code that need to
queuing discipline. In addition, links can be torn down or have these six types of information. Alternatively, using the
restored at any point in time duration the simulation, simu- NAM animator, user can design the model using Manu bar
lating link failures. and tools. In this case, the code(s) will be generated auto-
ns-3 has some shortcomings [26, 27]: matically in background for the user. The.ns file is a text file
and students can easily change/amend and learn about the
• The design of ns-3, using the hybrid of Tcl, otcl and C parameter(s) and, therefore, networking issues.
languages leads to memory consumption in many cases; In the node information section (Fig. 9), Node/site
• The IP address concept is not used; name(s), coordinates (x, y, z) of the nodes are codes. Link
• Alpha quality; section (Fig. 10) depicts the linkage between nodes/sites,
• Minimal documentation; link type (simplex, duplex), queue algorithm, etc. Agent
• Incomplete API; section (Fig. 11) covers the transmission protocol (TCP,
• ns-3 has no direct support for mobility or shared wireless UDP etc.), packet size, window size etc. Traffic section
radio channels.

SN Computer Science
SN Computer Science (2023) 4:269 Page 9 of 15 269

Fig. 8  Model processing steps in ns-3 showing how to create a tcl/.ns file and simulate in Nam staring with step 1a and ending at 3b

Network Model in ns‑3

Features of the network model in ns-3 are:

Node

1. The location of the node is in 3D view. That is, each

node has coordinates with the 3- dimensions (x, y, z).
However, the value of the z coordinate is always 0.
2. Node/site name must be one word (e.g. ‘London’, ‘USA’,
etc.).
3. The minimum coordinates of the nam (visualisation
tool) of the ns-3 are in the left-down corner and the
maximum are in the right top corner of the window (see
in Fig. 15).
4. Medium scale network simulation tool.

Link
Fig. 9  Node information of the.ns file showing the node name and
coordinates 1. The.ns formatted file supports only the simplex link,
whereas the.tcl formatted file supports both a simplex
and duplex link.
(Fig. 12) shows the application protocols, linkage with the 2. Visualisation of the node in ns-3 starts with 0, 1, 2… and
agents, time interval to pass the packets etc. Agent connec- so on. The first node is numbered 0 rather than 1.
tion section (Fig. 13) interlinks between the agents (source 3. Models designed by the users are not confined within
and sink). Finally Simulation Time section (Fig. 14) shows the predefined topological model/algorithm.
when the agent will start sending the traffic and how long 4. The notation of the.ns file (fixed style of notation for
simulation will continue. node name, agent name, traffic source name) and.tcl file
are slightly different. The.tcl file supports any text names

SN Computer Science
269 Page 10 of 15 SN Computer Science (2023) 4:269

Fig. 10  Link information of the.

ns file between the nodes, link
type

Fig. 12  Traffic information of the.ns file showing the application pro-

tocols

of a node, agent, or traffic source. On the contrary, the.

ns file supports only a particular format.

Fig. 11  Agent information of the.ns file showing the transport proto-

Traffic
cols

1. ns-3 supports several types of communication protocols.

2. ns-3 model can be animated by the network animator
nam.

SN Computer Science
SN Computer Science (2023) 4:269 Page 11 of 15 269

Figure 15 shows the visualisation screen of the nam tool.

The left lower side is the minimum points and right upper
corner is the maximum point/coordinate of the display.

BRITE

Brite [25] is a generator based on the AS power-laws. Fur-

thermore, BRITE also incorporates recent findings on the
origin of power-laws and observations of skewed network
Fig. 13  Agent Connection information of the.ns file showing the placement and locality in network connections on the Inter-
source and sink of the linkage net. By studying a number of existing topology generators,
the authors of BRITE claim that the preferential connec-
tivity and incremental growth are the primary reasons for
power-laws on the Internet. For completeness, topologies are
generated that incorporate both skewed node placement and
locality in network connections as well as topologies with
just incremental growth and preferential connectivity. To
generate a topology on a plane, the plane is first divided into
HSxHS squares, then the number of nodes in each square is
assigned according to the placement, NP which is either a
Fig. 14  Simulation Time information of the.ns file showing the start uniform random distribution or a bounded Pareto distribu-
of packet sending and duration of the simulation tion. The bounded Pareto distribution gives a skewed node
placement where a non-negligible number of squares have
a large number of nodes in them. Each square is further
divided into LSxLS smaller squares and the assigned nodes
are then uniformly distributed among the smaller squares. A
backbone node is selected from each of the top-level squares
populated with nodes and a spanning tree is formed among
the backbone nodes. Nodes are then connected one at a time
to nodes that are already connected to the backbone. A new
node can have preferential connectivity (PC) in its choice of
neighbouring nodes: locality based, outdegree based or both.
The locality-based preferential connectivity uses a Waxman
Fig. 15  Coordinate system of the tool Nam (GUI) animator showing
minimum and maximum points probability function to connect nodes in the topology. In
outdegree-based preferential connectivity, the probability of
a new node connecting to an existing node is the ratio of the
Visualisation existing node’s outdegree over the sum of all outdegrees of
nodes in the connected network. Finally, when mixing both
1. The network animator has a graphical editor in which locality-based and outdegree-based preferential connectiv-
locations of the nodes visualised are always the same. ity, the probability of connecting to an existing node under
However, during execution of the.nam file, this is not the outdegree-based preferential connectivity is weighted by the
case. Waxman probability between the new node and the existing
node. Each new node introduces new links.
Others
Files and Model Processing Steps in Brite
1. There is no performance analysis information defined
in the ns-3 model file as well as nothing is stored in the There are two text files in Brite whose processing steps are
trace files (.nam/.tr). shown in Fig. 16 below. The configuration file (text file) is
2. Produces a trace file (optional). The contents of the trace provided as input to the Brite and a model file is produced
file are of the same type as the.nam file. accordingly. Input can also be given graphically using the
3. No performance analysis is done. GUI of the tool. An input file can also be imported from
other tools (e.g. Inet). The student/user provide the con-
figuration parameter (GUI) or Topology file (from other

SN Computer Science
269 Page 12 of 15 SN Computer Science (2023) 4:269

Fig. 16  Model processing steps

in BRITE to design a model
starting with 1a and ending at 2

Fig. 17  The structure of the Brite file that shows the node and link information

topology generation tool) can be feed to Brite and Brite tool Network Model in Brite
general the.Brite file. The user analyses the file and will have
understanding of the topology concept. Features of the network model in Brite are:
The main file format that works behind different kinds Node
of network topologies in the Brite tool is shown in Fig. 17
as a sample [26]. This is basically the.brite file (output file) 1. The location of the node is 2D. That is, each node has
that can be produced manually or from a configuration file the coordinate with values of x and y. There is no z
(input). There are three sections that carry model informa- coordinate.
tion, node information (node number/label, coordinate) and 2. The nodes are labelled with 0, 1, 2, 3 and so on. The first
topology information (source, sink and some parameters) node starts with label 0 rather than 1.
respectively (details can be found in [26]). The configuration 3. The model does not work without the values of coordi-
file is not shown. nates.
4. Large-scale network topology generator.

SN Computer Science
SN Computer Science (2023) 4:269 Page 13 of 15 269

5. Generates a small backbone, with nodes placed: Others

6. randomly or
7. concentrated (skewed) 1. No performance analysis is done.
2. Contains eight different generation models: flat router
Link models (Router Waxman, Router Barabasi), Flat
Autonomous System (AS) models (AS Waxman, AS
1. The links of a model in Brite are duplex. Barabasi), Imported file models (BRITE 1.0, GT-ITM,
2. The new node has a constant number of edges connected NLANR, Inet, SKITTER), Hierarchical Top-Down
using: model, Hierarchical Bottom up model.
3. preferential connectivity and/or 3. Add nodes one at a time (incremental growth).
4. locality 4. To generate a topology on a plane, the plane is first
divided into HSxHS squares, then the number of nodes
Traffic in each square is assigned according to the node place-
ment (NP), which is either a uniform random distribu-
1. Brite does not support any communication protocol. tion or a bounded Pareto distribution. Each square is
2. No trace file is produced. further divided into LSxLS smaller squares.

Visualisation Two files are related to BRITE models: one configura-

tion file (input file), and one output file. These files are
1. It has no visualisation tool of its own. structured with a fixed format and keywords.
2. Node name, link name, etc. are denoted by numbers The Table 1 shows the main characteristics and comple-
rather than in text. mentary features of three selected tools. After extensive
3. No animation is supported. research on these education tools these features (summary)
have been found. Brite is used for LAN, MAN, WAN. On

Table 1  The main characteristics and complementary features of the Brite, Delite and NS tools
Criteria/ Tools Brite Delite ns-2/ns-3

Network topology generation Yes Yes No

Topology design method Some predefined models Predefined algorithms for automatic Manually designed by users
generation
Target network model LAN, MAN, WAN WAN WAN
Scale of network Very large scale Small-scale network Medium-scale network
Outputs Generated topology file Graphical- Link analysis, node utilisa- Simulation of several protocols
tion, overall utilisation in percent,
delay analysis
Model visualisation No Static Dynamic animation, Graphical Editor
Model data format Editable script file Editable script file Editable script file
Traffic model No In and out flow per node Comprehensive
Easy configuration of network param- No Yes Yes
eters
Application protocols No No CBR, FTP, HTTP, Telnet, Pareto
Transport protocols No No TCP, UDP
Routing protocols No No DSDV, DSR, AODV, TORA
Radio propagation No No Two ray, free space, shadowing
Radio model No No No Accumulating
Packet reception models No No No
Platform Unix Windows Unix, Windows
Open source code No No Yes
Implementation language C +  + , Java C C +  + and Otcl
Good documentation of the tool Yes No No
(users' manual)
Usages Educational, Research Educational Educational, Research

SN Computer Science
269 Page 14 of 15 SN Computer Science (2023) 4:269

the other hand, Delite and NS are used for WAN design. NS-3 can be used in the classroom. Finally to know the
Brite and Delite do not show the dynamic behaviour, network topology, BRITE tool case be used. We found
whereas NS is used to visualise the dynamic behaviour of one article regarding the tools for teaching network [3–7].
a network model. Please see the Table 1 for other features. The ns-3 (.ns) does not allow any other words rather
If these three education tools are used in teaching the com- than a set of predefined keywords. In Brite, the site name
puter networking course, students will be benefitted and (text) is not allowed and the site is indicated only by an
learn how to design and analyse a network model along id number. In contrast, Delite and .tcl (ns-3) allow a user-
with research. defined site name (text). In all cases, files are structured
in such a way that particular data must be placed in a fixed
column having spaces in a fixed manner. ns-3 simulates
Conclusions the dynamic behaviour of a protocol. It can simulate most
of the protocols from the TCP/IP family. Additionally, it
We have presented by examples the benefit of using mod- produces a trace file which can be analysed using the post-
elling and simulation tools in computer network courses. processing utilities. Another advantage is that the user can
We have shown the use of the network simulation and produce a model file using either a GUI or writing code
design tools for teaching, their main features, internal file (tcl script). The editor of ns-3 (nam) is very flexible for the
structure, format, relationship between files and file pro- naive as well as expert users. In addition, new protocols
cessing systems, how to design the network model and can be embedded into ns-3 and tested as the tool has open
learn, default features of the tools, etc. However, using source code.
hands-on projects, students will see themselves in differ- This article provides issues and techniques of three com-
ent (as designers); where, in those moments, they inte- puter network modelling and simulation tools for teaching
grate their minds, critical thinking, communication skills, Computer Networks. The theoretical concepts are rather
and their cumulative knowledge in the domain to provide abstract in networking subject, and students often find them
solutions to their problems. Moreover, this approach will too obscure to understand, many authors attempt to answer
enable faculty to notice and consider all issues facing stu- the question on how to improve students’ motivation and
dents when they apply the selected tools to solve their interest for the complex teaching material of computer net-
real-world like problems. working [7]. Our investigation suggests that the designs
“Pilot” tools for the teaching are Delite, network simu- could be started by building the network topology using
lator ns-3, topology generator Brite. We have shown the Brite, then gradually students are advised to add more capa-
technical issues how these tools are used to design a net- bilities and features to their designs such as adding links-
work model and the format of the input and output files performance, choosing right applications (using Delite), and
that a user/student needs to study to know/clarify the ideas add more advanced features such bandwidth, packet drop-
of the network models. To clear the idea of the topology, ping, various queuing mechanisms, build special network-
students will use the Brite tools. To analysis the network ing model between some sites in the network and quality of
models student can use the Delite tools and full analy- service (QOS) parameters to their designs (using NS).
sis details can be found in [14]. To visualise the dynamic
behaviours of network models (e.g. packet passing, packet
Data availability All the codes, models and required information are
loss, queuing, etc.), the NS-2/NS-3 tool would be very available on request.
beneficial.
Brite generates the small as well as very large topolo- Declarations
gies. It helps to design eight types of different network
models [17–26]. Also, models from some of the other tools Conflict of Interest The authors would like to declare that they have
no known competing financial interests or personal relationships that
(e.g. GT-IMT, Inet, Skitter, NLANR, etc.), can be fed into could have appeared to influence the work reported in this paper.
the Brite tool. Additionally, the output of Brite can be
transformed into SSF, J-Sim, Otter, ns formats. There are Open Access This article is licensed under a Creative Commons Attri-
great advantages of using this tool, i.e. the use of Brite bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
means use of the other compatible tools along with Brite. as you give appropriate credit to the original author(s) and the source,
Delite is a network modelling tool that has analysis provide a link to the Creative Commons licence, and indicate if changes
capability and predefined algorithms. Delite helps to were made. The images or other third party material in this article are
design and analyse a network model in addition to the uses included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
of the predefined algorithms. Delite tool can be used to the article's Creative Commons licence and your intended use is not
teach the calculation of the Capacity, reliability and delays permitted by statutory regulation or exceeds the permitted use, you will
of the networking module. To see the dynamic behaviours,

SN Computer Science
SN Computer Science (2023) 4:269 Page 15 of 15 269

need to obtain permission directly from the copyright holder. To view a 15. Noor NM, Yayao N, Sulaiman S. Effectiveness of using cisco
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. packet tracer as a learning tool: a case study of routing protocol.
Int J Inf Educn Technol. 2018;8(1):11–6.
16. Elias MS, Ali AZM. Survey on the challenges faced by the lec-
turers in using Packet Tracer simulation in computer networking
course. Proc Soc Behav Sci. 2014;131(2014):11–5.
References 17. Rocky KC, Chang, HK. Teaching computer networking with
the help of personal computer networks. Proceedings of the 9th
1. Atila B. Teaching computer networks: theory and problem solv- Annual SIGCSE conference on innovation and technology in com-
ing. J Adv Comput Netw. 2015;3(4):299–302. puter science educationIn, ITiCSE’04, Leeds, United Kingdom, p
2. Jiang D, Song H, Rong HW. Simulation tools and techniques 208–212, June 28–30, 2004.
for communications and networking. Mob Netw Appl. 2021. 18. Richard VS. Network planning and design. Polytechnic University,
https://​doi.​org/​10.​1007/​s11036-​021-​01746-0. New York. Revised 21 Jan 2001. https://​study​lib.​net/​doc/​14653​
3. Vandana K. A study of research tools and techniques in net- 654/n​ etwor​ k-p​ lanni​ ng-a​ nd-d​ esign-b​ y-r​ ichar​ d-v​ an-s​ lyke-p​ olyte​ ch.
work congestion control. Int J Eng Res Technol (IJERT). Accessed 2 Mar 2023.
2014;3(2):6–10. 19. Arnold WB. Which network design tool is right for you? IT Pro-
4. Yongbin Z, Ronghua L, Huiling M. teaching innovation in com- fessional 2000;2:23–31. https://​doi.​org/​10.​1109/​6294.​877494
puter network course for undergraduate students with packet 20. Rahman MA, Algirdas P, Frank ZW. Network modelling and
tracer. IERI Proce. 2012;2(2012):p504-510. simulation tools, simulation modelling practice and theory. Int J
5. Antonio R, Fernando P, Rafael M, Pedro MR, Antonio Fed Eur Simul Soc EUROSIM. 2009;17(6):1011–31.
FS. Teaching advanced concepts in computer networks: 21. Rahman MA, Algirdas P, Frank ZW. An integrated environ-
VNUML-UM virtualization tool. IEEE Trans Learn Technol. ment for network design and simulation. In: Sobh T, Elleithy K,
2013;6(1):85–96. Mahmood A, Karim M, editors. Advances in computer, informa-
6. Marina P, Julije O. Methods in teaching computer networks: a tion, and systems sciences, and engineering. Springer Publication;
literature review. ACM Trans Comput Educ. 2020;20(3):1–35. 2006. p. 316–22.
https://​doi.​org/​10.​1145/​33949​63. (Article No.: 19). 22. Cahn RS. Wide area network design: concepts and tools for opti-
7. Azeddien MS, Mohamed J. Using simulation and modeling tools mization. San Francisco: Morgan Kaufman Publishers; 1998.
in teaching computer network courses. In: International Confer- 23. Kevin F, Kannan V. The ns Manual. The VINT Project, 23 May
ence on IT Convergence and Security (ICITCS), Macao, China, 2006. http://​www.​isi.​edu/​nsnam/​ns/​doc/​ns_​doc.​pdf. Accessed 2
6–18 Dec 2013. Mar 2023.
8. Perc M. High-performance parallel computing in the classroom 24. Deborah E, Mark H. Network Visualization with Nam, the VINT
using the public goods game as an example. Eur J Phys. 2017;38: Network Animator Computer 33(11):p63–68, 2000. https://​doi.​
045801. org/​10.​1109/2.​881696, Source: IEEE Xplore.
9. Silva PHO, et al. Graphical interface as a teaching aid for nonlin- 25. Alberto M, Anukool L, Ibrahim M, John B. BRITE: an approach
ear dynamical systems. Eur J Phys. 2018;39:065105. to universal topology generation. modeling, analysis and simula-
10. Arnold W. Teaching complex ideas: how to translate your exper- tion of computer and telecommunication systems, 2001.In: Pro-
tise into great instruction. New York: Routledge; 2019. (ISBN: ceedings. Ninth International Symposium, Feb 2001. https://​doi.​
9781351058117). org/​10.​1109/​MASCOT.​2001.​948886.
11. Pakstas A. Reliability issues in network planning. In: Proceed- 26. Riley GF. The Georgia tech network simulator. In: Proceedings of
ings of the 8th International Conference on Telecommunications, the ACM SIGCMM 2003 Workshops, Germany, 2003, pp 5–12,
ConTEL, Zagreb, Croatia, 15-17 June, 2005. Aug. 2003. https://​doi.​org/​10.​1145/​944773.​944775. Accessed 2
12. Ping W, Raed S. Using a team project in teaching com- Mar 2023.
puter networking and data communication. Issues Inf Syst. 27. Srivastava M, Partho M, Prathima A, Nguyen G. Ethersim: a
2020;21(2):167–77. simulator for application-level performance modeling of wire-
13. Rashid NA, Othman Z, Johan R, Sidek SFH. Cisco Packet Tracer less and mobile ATM networks. Comput Netw ISDN Syst.
simulation as effective pedagogy in computer networking course. 1998;29:2067–90.
Int J Interact Mob Technol (iJIM). 2019;13(10):4–18. https://​doi.​
org/​10.​3991/​ijim.​v13i10.​11283.-5. Publisher's Note Springer Nature remains neutral with regard to
14. Zhamanov A, Yoo S, Sakhyeva Z, Zhaparov M. Implementation jurisdictional claims in published maps and institutional affiliations.
and evaluation of flipped classroom as IoT element into learning
process of computer network education. Int J Inf Commun Tech-
nol Educ. 2018;14(2):30–47.

SN Computer Science