The performance of ATM network

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.



Although there are many local and wide area network technologies still in use, research and development is still going on. IP and ATM are the most widely used technologies and are compatible.

IP (Internet protocol) is a well known and commonly used internet technology it is a very important protocol in internetworking. The current version of internet protocol (IPV4) uses only 32 bits , the number reduced drastically however a new version of internet protocol (IPV6) has been has been introduced the second version of internet protocol to be used, this can hold up to , (3.4 x ). Research has not stopped here MPLS (Multi protocol label switching) has come up also as an IP network technology. This is a connectionless protocol and cannot keep track of the conversation.

ATM (Asynchronous Transfer Mode) is a packet switching protocol that encodes data into small fixed size cells. It deals mainly with real-time video conferencing and audio as well as image files, text and e-mail. It establishes a virtual circuit between end points before the actual data exchange begins.

This project deals with comparing the performance of IP and ATM network characteristic and behaviour of real-time video, audio, image, text, and e-mail files traffic. About eight scenarios are built to perform this comparison study.

The aim of this project is to find out the best technology with respect of delay, jitter, end-to-end delay, link utilisation in audio and video data packets, and also to study about the working style of the technology and its dealing with data traffic.


CHAPTER 2: Literature review: where the back ground discussion is done before starting the project.

CHAPTER 3: Network modelling: Discuss issues of network modelling and its benefit.

CHAPTER 4: Company Network model: Discuss in detail the two network model.

CHAPTER5: Network simulation and analysis.




During the 20th century, the key technology is the information gathering and distribution. As a result of rapid technological progress in the computer industry, the main goal of computer network is to share information rather data, store, and process information over a wide geographical area at a push of a button. As our ability to gather process and distribute information, the demand for even more sophisticated information processing grows even faster.

Network is not a new word. This simply means an interconnected systems of things or people [web dictionary] there are so many types networks present. E.g. a bus network, radio network, television broad cast network and telephone network but all these networks are completely different with modern technology of computer networks.

Although computer industry is still young compared to other industries (e.g. automotive and air transportation), computers have made a spectacular progress in a short time and have changed the whole concept and life style of the world and provides more flexibility to do work. In a simplest form a computer network can be defined as: A computer network is a group of inter connected computers linked in order to share resources (such as printers and CD-ROMs), exchange files, or allow electronic communication. Computers on a network may be linked through cables, telephone lines, radio waves, satellites, or infrared light beams.[]

Computer usage has excessively grown in every aspect of life for example: business, research, education, government etc. Then the development of "INTERNET" which is defined as the global system of interconnected computer networks that use the standardized Internet Protocol Suite (TCP/IP). It is a "network of networks" that consists of millions of private and public, academic, business, and government networks of local to global scope that are linked by copper wires, fiber-optic cables, wireless connections, and other technologies, it gives an idea of the whole world as a global village. The internet has become a vital part of every organization, business or individuals to do business (must have internet connection). As compared to old networks, individuals were unable to access networks of others due to incompatibility issues with the software and hardware used. This incompatibility has been resolved by the introduction of networking software such as "Middleware" to allow them to communicate with each other. Now organizations and individuals can exchange and transfer data (information) across networks in less time and accuracy.


Architecture is the basic building block of any structure; it has been informally considered to consist of a frame work, some components, and rules. A communication system has to define some rules, functions, and operation to build its architecture. There are two industry standard network architectures:

a) TCP/IP (Transport Control Protocol Internet Protocol).

b) OSI (Open System Interconnection).

Before the introduction of these standard communication protocols communication between applications was tightly dependent on particular specific programs making it difficult for one vendor application to communicate with another vender application which created problems for the end-user to choose which application to use to do their daily work. After introducing these standards life became easier for communicating devices, doesn't matter who is the manufacture of the device because this architecture implements certain communication protocols in devices for the ease of not only their user but also to make international standards for communications which is regulated by IETF (Internet Engineering Task Force).

In discussing computer networks, two concepts are paramount:

§ Protocols

§ Computer architecture, or protocol architecture

For two computers to communicate successfully, they must "speak the same language." What is communicated, how it is communicated, and when it is communicated must conform to some mutually acceptable conventions between the entities involved. The conventions are referred to as a protocol, which is defined as a set of rules governing the exchange of data between two computers. [Data and Computer Communications Sixth edition William Sterling page chapter one 13]

As we discuss the key features of protocols, the two most important protocol architectures are: the open system interconnection (OSI) model and the TCP/IP model. Although it's the TCP/IP protocol suite that is the basis for most commercially available interoperable product.


Transmission Control Protocol / Internet Protocol, is a communication protocol suite which has served the network traffic for decades, it came as an outcome of the research committee called ARPANET (Advanced Research Project Agency Network) sponsored and under the supervision of the DoD (U.S. Department of Defence) and was the first wide area packet switching network which evolved into the INTERNET.

TCP/IP is a combination of two protocols TCP and IP, this protocol defines the way of exchange of information between two or more networks. TCP is a connection oriented and IP is a connectionless oriented protocol, combining them gives a reliable communication flow between different network hosts. TCP operates on a transport layer and IP operates on a network layer of the OSI reference model. It is a layered architecture having four (4) layers namely:

  1. Application
  2. Transport
  3. Internetworking
  4. Network access

TCP was developed before OSI model, they both have deferent number of layers, but they have been designed in a way that they can both communicate and understand each other at deferent levels whereby TCP/IP is important for addressing and routing data packets over the internet.

The application layer of TCP/IP model covers the application, presentation, and session layers of the OSI and the Network access layer covers physical and data link layers of the OSI model as shown in the figure below.

Comparison between (a) TCP/IP and (b) OSI model

§ Application layer:

This layer contains all the high-level protocols, these include TELNET (virtual terminal), which allows users on one machine to log onto a distant machine and work there. FTP (file transfer protocol), provides a way to move data efficiently from one machine to another. HTTP (Hypertext Transfer Protocol) provides means of fetching pages on the World Wide Web, SMTP (Simple Mail Transfer Protocol), under the TCP protocol stack it covers the Application, Presentation, and Session layers of the OSI model. It is a layer that deals with user application.

§ Transport layer:

It is a layer above the internet designed to allow peers on the source and destination to carry out a conversation, just like in the OSI transport layer. Two end-to-end transport protocols have been defined here. The TCP (Transmission Control Protocol), is a reliable connection-oriented protocol that that allows a byte stream originating from one machine to be delivered without error on any other machine on the internet. The second protocol in this layer is UDP (User Datagram Protocol), is an un reliable, connectionless protocol for applications that do not want TCP's sequencing of flow control and wish to provide their own. It is used for client-server-type request-reply queries and applications in which prompt delivery is more important than accurate delivery, such as transmitting speech or video.

§ Internet layer:

The internet layer defines an official packet format and protocol called IP (Internet Protocol), it is a connectionless protocol used to deliver IP packets to their destination. Packet routing is the major issue here controlled by routers which route packets between network devices and nodes. This protocol provides logical addresses for the identification of source and destination. Data packets on this layer can be routed following different paths on the network and may reach their destination in a different sequence, re-ordering of this data packets is done by TCP.

§ Network layer:

This layer is responsible for the logical interfacing between an end system and a network; it combines the features of physical and data link layer of the OSI model. At this layer, the computer system engages in a dialogue with the network to specify the destination address and to request certain network facilities, such as priority. [RFC 1122]


The Open System Interconnection Model is a logical layered frame work developed by the international organisation called the ISO (International Standard Organisation), it deals with connecting open systems i.e. systems that are open for communication with other systems.

"An open system is a set of protocols that allow any two different systems to communicate regardless of their underlying architecture". [Behrouz Forouzan, Cathering Coombs and Sophia Chung Fegan, 1998, Chapter 3: the OSI Model, page 35]

The OSI model has seven layers as shown in figure 2.1 namely: Physical, Data Link, Network, Transport, Session, Presentation, and Application from which each has its own function.

The OSI reference model layers are divided into subgroups, we can group Application, Presentation, and Session layers together because they interact with users directly and handle the software part. Network support group includes Physical, Data Link, and Network layer which are physical and hardware part of the communication system e.g. addressing, sending data and establishing communication, while the third group which is Transport is use to join these two other groups with each other. Every layer accept Application and Physical adds some control information to the header of the data unit and passes this PDU (Protocol Data Unit) to the next layer. Communication between layers is made possible by the interfaces which decide what information and services should be passed on to the next layer. Local information and services are kept hidden from other layers. On the other hand the receiving side, the control information is subtracted from the PDU to extract the information part from it. This layer modularity communication gives scalability to make changes in any layer without disturbing the functionality of other layers. [Computer Networks fourth edition Andrew S. Tanenbaum].

OSI Model Layers

Seven layer OSI Model

§ Application Layer:

It contains all protocols and methods that fall into the category of process to process communication via the internet protocol (IP) network using the transport layer protocol to establish communication.

§ Presentation Layer:

This layer provides strict modular separation of functionality and provides protocol implementation for each layer.

§ Session Layer:

Session layer provides a mechanism for opening, closing and managing a session between end-user application processes.

§ Transport Layer:

This is a group of methods protocols with in a layered architecture of network components within which it is responsible for encapsulating application data blocks into data units (Datagram and Segments) suitable for transfer to the network infrastructure ready for transmission to the destination host.

§ Network Layer:

It is responsible for end-to-end (source to destination) packet delivery including routing through intermediate hosts.

§ Data link Layer:

It is a protocol layer which transfers data between adjacent network nodes in a wide area network or between nodes on the same local area networks segment.

§ Physical Layer:

This layer is the first and lowest layer in the seven layer OSI model. It consists of the basic hardware and transmission technologies of the network.


There are mainly three types of network architectures namely:

a) LAN (Local Area Network).

b) MAN (Metropolitan Area Network).

c) WAN (Wide Area Network).

Classification of interconnected processors by scale [Computer Networks by Andrew S .Tanenbaum Fourth edition page 16]


LAN are privately own computer networks which supply networking capability to a group of computers in a close proximity to each other such as in an office building, a University or a home. A LAN is useful for sharing resources like files, printers, games or other applications. A LAN in turn connects to other LANs, and to the internet or other WAN. It has a low cost implementation as compared to MAN and WAN. [Computer Networks by Andrew S .Tanenbaum Fourth edition page 16]

LANs are restricted in size, and use a transmission technology consisting of a cable to which all the machines are attached. LANs run at a speed of 10Mbps to 100Mbps and have a low delay (microseconds or nanoseconds), and make very few errors.


IEEE 802 (Institute of Electrical and Electronic Engineering) is a standard for interconnection between network devices over the network. Among these standards ETHERNET 802.3 is commonly used for LAN networks, which particularly specifies the functions of physical layer and data link layer of the OSI (Open System Interconnection) model. OSI model is the reference model for internetworking designed and defined by ISO (International Organisation for Standardisation). [Behrouz Forouzan, Cathering , Coombs and Sophia Chung Fegan 1998].

A list of Ethernet Standard

i. Wireless LAN (802.11).

ii. Wireless Personal Area Network (802.15).

iii. Token Ring Access Method (802.5).

iv. Logical Link Control (802.2).

v. Broadband Wireless Metropolitan Area Network (802.16).


IEE LAN Standards compared with OSI model [Alberto, Leon-Garcia Indra Widjaja, 2000].

The relationship between OSI reference model and IEEE 802 is shown in the figure above. Here, the IEEE 802 committee has subdivided the DATA LINK layer of OSI model into two sub-layers.

a) LLC (Logical Link Control) upper sub-layer of the data link layer.

b) MAC (Media Access Control Layer) lower sub-layer of the data link layer.

LLC layer is responsible for controlling the framing error and sequencing the frames. It is architecture independent, which means it is common in all LANs defined by the IEEE 802. Also it provides a way of exchanging frames between LANs which uses different MAC protocols. Data unit at this layer is known as PDU (Protocol Data Unit), which includes DSAP (Destination Service Access Point), SSAP (Source Service Access Point), control information and data. Bear in mind that MAC layer address is fixed, so this logical address (Service Access Points) can distinguish between multiple data of different layer protocols used by the workstation at any given time for example checking emails, listening to songs, reading news papers etc.

MAC (Media Access Control) layer has a fixed physical address of 48 bits which uniquely identifies each computer in the world. The starting 24bits are reserved from OUI (Organisational Unit Identifier) and the last 24bits are reserved for vendors. MAC address helps to recognise a workstation over the network to enable communication with each other. As you can see in figure 2.1, that this layer has several MAC standards like (802.3, 802.11) and each standard has its own physical medium of flow of data. It mainly deals with the access method of shared physical media in between the network users. On this level PDU contains source and destination physical address (MAC address), checksum which uses an algorithm of CRC (Cyclic Redundancy Check). It controls the order delivery of frames and notifies if an error occurs. LLC frame is encapsulated I MAC frame as shown in figure 2.2. [Computer Networks Fourth Edition Andrew S. Tanenbaum].

LLC and MAC frame PDU Encapsulation. [Computer Networks Fourth Edition Andrew S. Tanenbaum page 291].

IEEE 802.3 (CSMA/CD)

The most commonly used medium access control technique for bus and star topology is carrier sense multiple access with collision detection (CSMA/CD). This mechanism allows multiple workstations of the network to share the transmission medium at the same time, and due to this mechanism a collision may occur over the network between data of two different workstations. To avoid this, collision detection (CD) is used to improve CSMA performance by terminating transmission as soon as a collision is detected, and reducing the probability of a second collision on retry.

To fully understand the concept of CSMA/CD, Imagine a very simple Ethernet network with only two nodes. Each node, independently, decides to send an Ethernet frame to the other node, and then both nodes listen to the Ethernet wire and sense that no carrier is present. Both nodes transmit simultaneously, causing a collision. Both nodes detect the collision and each node waits a random amount of time before transmitting again.

By using CSMA/CD technique in LAN network, nodes can communicate with minimum risk of collision and therefore maximise data throughput on the medium.


SFD = start of frame delimiter

DA = destination Address

SA = source Address

FCS = frame check sequence

IEEE 802.3 frame format

§ Preamble: A 7-Octet pattern of alternating 0s and 1s (10101010) used by the receiver to establish bit synchronisation.

§ Start of frame delimiter (SFD): It has a 1 byte field having a similar bit pattern of preamble field. The sequence 10101011 which indicates the actual start of the frame and enables the receiver to locate the first bit of the rest of the frame.

§ Destination address (DA): It is a 6-byte field which specifies the station(s) for which the frame is intended. It may be a unique physical address, a group address, or a global address.

§ Source address (SA): It is also a 6 byte field, specifies the station that sent the frame.

§ Length/type: It is 2-byte specifies the length of LLC data field in octets, or Ethernet type field, depending on whether the frames conforms to the IEEE 802.3 standard or the earlier Ethernet specification. In either case, the maximum frame size, excluding the preamble, SFD, is 1518 octets.

§ LLC data: 2-byte field which indicates the Data unit supplied by LLC.

§ Pad: Octets added to ensure that the frame is long enough for proper CD operation.

§ Frame check sequence (FCS): A 32-bit cyclic redundancy check, based on all fields except preamble, SFD and FCS. [Data & Computer Communications Sixth Edition by William Stallings, page 475].


LAN Topology

Topology or network topology means the physical or logical arrangement of the elements of a network. During the design stage of the virtual arrangement of the network scalability has to be considered for the network to have the ability to expand as additional users may be required for different functionality. Below is a brief summary of the three basic types of topologies.

i. Bus topology

ii. Ring topology

iii. Star topology

i. Bus Topology

It is one of the basic types of topology in which all nodes are connected together using a single bus.

Bus topology

As shown in figure 2.5 the structure of a bus network all workstations are attached appropriately and directly to the transmission medium, or bus. A transmission from any workstation propagates the length of the medium in both directions and can be received by all other work stations. At each end of the bus is a terminator, which absorbs any signal removing it from the bus.

ii. Ring Topology

Ring as the name implies is a network topology where computers are arranged in a circular format making a closed loop as shown in figure 2.6. In this ring topology links are unidirectional; that is, data are transmitted in one direction only and all are oriented in the same way.

Ring Topology

iii. Star Topology

In this Star topology, each workstation is directly connected to a central node called "hub /switch". Typically each workstation attaches to a central node via two point-to-point links, one for transmission and one for reception. The biggest advantage of star topology of connecting each workstation individually is robustness, which means if one workstation or single link is down other workstations, can still communicate. It is also easy to trace and narrow down the fault. I have used this topology not only because of its advantages, but also easy to implement and configure.

Star topology.


A MAN is optimised for a larger geographical area than a LAN, ranging from several blocks of buildings spanning to a city. MANs can also depend on communication channels of moderate-to-high data rates. A MAN can be operated, owned by a single organisation, or operated as a public utility. (For example the Metropolitan Police of United Kingdom and cable Television).

Metropolitan Area Network

These days this technology is phasing out because of the new technologies like the VPN (Virtual Private Network), Intranet. This is archived by forming links between nodes over logical connections or virtual circuits between hosts of larger network such as the internet. The link layer protocols of the virtual network are said to be tunnelled through the transport network.


A wide area network, or a WAN, spans a large geographical area, often a country or a continent. It contains a collection of machines called hosts which are connected by a communication subnet. Hosts are owned by the customers (e.g. people's personal computers) whereas the communication subnet is typically owned and operated by a telephone company or Internet service provider. The subnets carry massages from host to host and data travels long distances at a high speed in high data rate transmission lines. Technologies like ATM (Asynchronous Transfer Mode), Frame Relay, X.25, ISDN, T1/T3 lines are required. [Computer Networks Fourth Edition by, Andrew S. Tanenbaum].

Among several Technologies my discussion and focus in this project is on two of them (a)IP (Internet Protocol) and (b) ATM (Asynchronous Transfer Mode).


The internet protocol (IP) is part of the TCP/IP suite and is the most widely used internetworking protocol. It is unreliable connectionless protocol used for transmission of data, operated on a network layer of OSI reference model. For reliable connection and transmission of data between end-to-end systems Transmission Control Protocol (TCP) which operates on transport layer of the OSI reference model is used in conjunction with IP in order to archive a reliable connection. As discussed earlier TCP/IP architecture compared with OSI model, TCP is a connection oriented protocol which uses a three-way hand shake for the successful transmission of data.

The following are the steps of the three-way hand shake:

  • The sender sends its initial sequence number (SYN bit) and the port to which it will use to communicate with the target host along with the window size information. This is also known as session creation.
  • The target host acknowledges the host's initial sequence number and sends its own sequence number to the host machine.
  • The sender acknowledges and receives the target's initial sequence number.


As discussed earlier, the TCP operates on transport layer of the OSI reference model and data unit on the layer is known as data segment. With respect of application point of view TCP provides the several services written below. [Douglas E. Comer, 2001]

a) Connection Orientation

b) Point-to-point communication.

c) Complete reliability.

d) Full duplex communication

e) Stream interface.

f) Reliable connections start up.

g) Graceful connection shutdown.



Network modelling is the very first step involved in network planning, design and implementation. Once design correctly, it helps to analyse the strength, stability and performance of the network on different types of data in this case (real time traffic and non-real time traffic), on different timings (peak and off-peak), and when something goes wrong (any device goes down) and how the model can cope with it.

"A model is an abstract representation of an item or concept .e.g. plane or building or part of something. Models are created in order to view, manipulate or test the thing they represent without having to build the real thing". [Bernad Bruegge & Allen H. Dutoit, 2004]

Models are useful when dealing with systems that are too large, too complicated or too expensive to experience first time. Models also help us to visualise and understand systems that either no longer exists or that are only claimed to exist.

Network analysis and designing are the steps before network modelling. The analysis phase defines the requirements, goals and problem areas of the network going to be built. Collecting information and types of traffic and their characteristics such as data rate, traffic types are the key points at this stage. Network implementation and cost are the highest point in any analysis phase. After that, network managers proceed with the design phase where they decide appropriate network topology, policies and communication procedures between entities of the network.

Modelling has a major role in developing either a new system or upgrading the existing system, by implementing pilot project and running current system can actually give a good idea of the changes made. By Making changes in the system according to the requirements and comparing them with the old system is the key solution to get success.

Network modelling helps a lot and gives enough time to think over more and more ideas, over worse conditions that could be possibly occur and to select the best possible solution which could be implemented out of the possible scenarios.


Among many network modelling tools OPNET (Optimised Network Engineering Tools) is one which has vast variety of network types, protocols, devices and up to date technology for the ease of usage of the end user and it is designed with a friendly graphical user interface.

Some features of OPNET Modeller 9.1

Opnet modeller has many features including:

§ Project Editor

Contains network nodes, connection links used to connect network models. Drag and drop facility is available to put instance of the object from the object palette in the work space, rapid configuration for quick and easy use of network topology creation. The geographical representation of devices eases the load of designing network topologies.

§ Node Editor

This gives a clear picture of the inside architecture of a node, by examining the dataflow inside that object and between the objects. Nodes can send, receive and generate network application traffic to each other.

§ Process Editor

This is mainly to do with the processes and events generated due to the execution of certain programs, applications or processes running over the network. During simulation every node may generate a process either by its self or according to any other events that may happen over the network, so it tells the state of the process and its behaviour.

Simulating a net work cannot be compared to real time traffic of the real world; it can only give an idea or prediction about certain aspects, such as:

§ How much bandwidth will be required for sending traffic or if some other traffic is added to it and in future what will happen?

§ Functioning of protocols and devices in the worst scenario and the scalability of the network

§ What happens if congestion occurs and if how will it be handled?

Traffic Load on Network

To get the maximum output of the simulation results one should create the maximum load of the application. By doing this the stability of the network and functionality of device can be examined. Application configuration in Opnet modeller is the object where network applications are defined together with their configurable options and parameters such as; send / receive interval time, size of bytes, types of service. Profile configuration groups the applications under a single object which eases the operation of assigning every individual application to every single node on which application(s) are assigned automatically and each node will link to those application(s) and generate the traffic.


It plays an important role in the simulation process. This is a random number where the sequence starts to generate the network traffic. There are two types of random numbers:

a) Pseudo Random: where the numbers are not repeated during the whole simulation time and are truly random sequence.

b) Seed Random: where numbers may be repeated in the sequence during simulation.



Two enterprise models are connected with each other using internet protocol (IP) and asynchronous transfer mode (ATM) as a back bone technology to help individually study the behaviour of different types of traffic such as:

a) Email

b) FTP (file transfer)

c) Database

d) Video

e) Audio

Opnet modeller 9.1 software is used for simulating and designing these models.

One enterprise model is taken from an Advertising Agency Eros International's marketing department (MK_DEPT) in London and their Production Lab in Reading (PROD_DEPT).

Before any further discussion, basic architecture of both Departments should be explained for better understanding of the networks.


Below is the diagram of MK_DEPT designed and configured using Opnet modeller 9.1. In reality there are to subnets of 15 computers each, but for the purpose of generating more traffic load over the network, one more subnet is added to it with the same number of computers.

Star topology concept is used to connect all three local subnet switches to the single central switch using 100-BaseT link. Two servers (email server and ftp server) are connected directly to the central switch also using 100-BaseT link providing services of applications running over the network. Three basic applications are configured using application configuration object; Email (heavy), Email (light), and Ftp (heavy) and are assigned in a single profile configuration. In order to obtain and make a comparison a base line and initial simulation will be taken and at this stage no video and audio traffic is added to both MK_DEPT and PROD_DEPT, separate scenarios will be take for these type of traffic later.


Figure 4.2 shows the complete network diagram of the PROD_DEPT model having labs in five subnets and two servers for the application services, web server (HTTP) and database server. Switch to switch links are 100-BaseT links and servers to switch links are also 100-BaseT.

PROD_DEPT Network model

Database (heavy), Database (light), Http (heavy), Http (light), are the applications configured in the application configuration object for the network in three separate profiles.




These are the names of the profiles in which applications are assigned.

Profiles of the second scenario

PROD_GEN = Database (light), Http (light).

PROD_MM = Http (heavy), Http (light), Database (light).

PROD_DB = Database (heavy), Http (light), Database (light).

Subnet 1 and Subnet 2

Two separate general purpose labs are implemented in two subnets (1 & 2) individually.

Figure 4.3 shows their implementation. A single group of twenty five workstations in a model of star topology in subnet_1 and two groups of twenty five workstations in the same model each group is connected through a local central switch and then to the main central switch. PROD_GEN profile is assigned to both subnets.

Subnet 3

This lab (lab3_Multmedia) is for Audio / Video conferencing, it also has the facility of normal applications traffic. The setup of the network subnet is the same as the previous one except the three subnets are grouped in a single subnet instead of being in separate subnets. All three central local switches are connected in star topology design to the central switch and then the central local switch is connected to the main central switch as shown in the figure below. Video conferencing is available here.

Subnet_4 and Subnet_5

These two subnets (lab_4 and_lab_5) are setup particularly for database applications but web browsing is also available. Both subnets have two groups of twenty five workstations connected directly with a switch in star topology. One central local switch connects the two switches of each group in the main central switch. PROD_DB is configured on all work stations. See figure 4.5.

The ease of use of Opnet modeler 9.1 Software gives vast variety of network designing technologies and devices for implementation. Drag and drop, facility from the object palette makes the work much faster for the end user not only that the profile assigning task and creating multiple subnets with the same configuration is also useful. Rapid configuration and duplicate scenarios are the two most important features of it through which end users can make a choice of any network topology. Selection of devices and links is also fast through drag and drop from the object palette.

Hence the basic architecture of each Enterprise model has been defined and explained. The next step is to combine them in different scenarios for multiple simulation using IP and ATM cloud and examine them by taking results of the required parameters. A list of possible scenarios is written below



List of scenarios.

Scenario 1 and 2 are typical base line scenarios for examining and testing the initial traffic. Basically, in these scenarios both MK_DEPT and PROD_DEPT are separated and checked the network performance and stability by running basic network traffic over them.

Both networks first time ever are joined together in scenario 3, by using IP cloud as back bone technology shown in figure 4.6. Two routers are used to connect both networks with each other. A 1000

Base-T link from central switch to router is used on both sides of the network for the connection. Point-to-point (PPP) connection of DS3 data rate link is used for the connectivity between routers to IP cloud.

By making two advancements in the scenario 3, scenario 5 is created. First, implementing and configuring 'quality of service' object (see appendix 'A' 'how to implement Quality of Service') secondly, removing FTP application server from MK_DEPT network and adding this service to the PROD web server as more end users are there as compare to number of MK_DEPT network end users. See figure 4.7.

In between scenario 3 and 5 there is scenario 4 which tests the changing effect of FTP application server. The changing effect results are achieved by successful examining 'FTP download response time' and 'FTP link utilization' between the server and the central switch.

After that scenario 6 is implemented using the second hand back bone technology ATM cloud to record the same parameter fields as recorded in IP network for further comparisons and judgments shown in figure 4.8, For the purpose of comparing with the best IP network (having QoS) there is no implementation and configuration of QoS object in this scenario. The aim is to give full strength of IP network with maximum good performance parameters and then compare those reading with ATM network reading. See figure 4.8.

In last two scenarios audio and video traffic is added with normal data traffic. Scenario 7 as shown in figure 4.9, audio and video traffic is added along with quality of services in terms of type of services (best effort, excellent effort, interactive multimedia, and interactive voice ) and weighted fair queue.

One subnet of MK_DEPT is reserved for generating audio and video traffic with normal data traffic. Both multimedia communications are in multipoint design which means one audio client is talking to many audio clients of far end network PROD_DEPT, and vice versa for audio clients.

Scenario 8 is similar to last scenario with ATM back bone technology having the same multimedia traffic.

Below figure 4.10 is the lab_3 of PROD_DEPT network which shows the multimedia traffic clients. Seven from each traffic so all together fourteen clients are there.

PROD_DEPT lab3 (multimedia)


It is one of the features of Opnet modeler to provide the facility to the network administrators or end users to define applications used over the network in an object called 'Application Configuration'. This Application Configuration has an object-oriented technique just like a procedure or function defined within a class can be called as many times from outside the class as required. It eases the work load by providing a central controlling point where all parameters of any specific application can be configured. Figure 4.11shows the list of applications used in the network. See appendix B for detailed description of each application parameters.


Applications are grouped together in a single entity. Profiles are assigned to network devices (servers and workstations) for generating network traffic of those applications which are included in that profile. Four profiles are created in all scenarios except the last two in which audio and video traffic is added in their respective profiles. See figure 4. 12.

Some parameters can be seen in figure 4.12 like operation mode (tells the mode of execution of the profile), starting time (when the profile starts or generate the traffic and on which mode) it has been given constant of (50) which means that after exactly 50 seconds this profile will be executed and will start to generate network traffic of the assigned applications. Duration (time it will take to run), repeatability (how many times it should the traffic should repeat its self). See appendix 'C' for details about profile configuration.



This project is started from the base line simulation of individual network (MK_DEPT, and PROD_DEPT). The purpose of doing this is to analyse the behaviour of basic data traffic over the network and to test the configuration files and make the necessary changes in the traffic characteristics and parameters.


This is the first base line scenario of the network as described earlier, email (heavy, light), and ftp (heavy), are the services configured in this scenario.

i. Email and Ftp Download Response Time

As shown in figure 5.1 the download response timings of email and ftp are 34 milliseconds and 146 milliseconds respectively. They are quite less values because of less number of users and it is a small scale network.

Scenario 1 Email Baseline reading before generating Traffic Load on the Network

ii. Email and Ftp Server Load

The figure 5.2 shows the email server load in respect to traffic received bytes/sec. the email traffic (heavy) is 77036.96 bytes/sec and email traffic (light) is 18042.27 bytes/sec.

Average Ethernet speed per second

Average Ethernet delay per second

This figure shows the entire Ethernet delay of the whole network in respect to traffic received Email and Ftp are 56 µ seconds.

Figure shows a similar parameter with a different reading of Ftp server load. Only heavy Ftp traffic is selected for this scenario so there is no Ftp (light). Ftp server load traffic received bytes/sec is 138867.3

It is very clear from the reading that Ftp traffic is much heavier than email traffic. Ftp server load is in six (6) figures while email server traffic load is in five (5) figures.


This is the second base line scenario of the second network model. Database (heavy, light) and Http (heavy, light) also applications are configured in this network.

The key element in this scenario for future references is Http page response time; other elements are taken into consideration as well.

i. Database and Http Server Load

At this stage Database and Http heavy server load with respect to traffic received bytes/sec is shown as a point to see the stability of the network with maximum load. The server load with heavy traffic of Database and Http is 8338.28 bytes/sec and 9000 bytes/sec respectively.

Database Server Load (Heavy)

Http Server Load (Heavy)


Here, two scenarios are directly compared to see the real effect of changes in the scenarios. As both scenarios are similar except the one configured with the QoS (Quality of Service) parameter. The purpose of implementing QoS is to record the behavior of network traffic with and without it.

New parameter sand devices are set and implemented time to time to make the comparison effective with the base line network.

ii. Central link Utilisation (%)

A clear comparison can be seen in figure 5.5 between link utilisation with and without QoS. Without QoS the central link utilisation is 14.7% and with QoS, its utilisation increases up to 8.54%. It does not only increase but also stabilises as more curst and curves are available which shows signs of heavy utilisation of the link.



Central link utilisation (%) comparison.

iii. Queuing Delay

Figure 5.6 clearly shows the queuing delay between the above two scenarios. The upper pane of the figure belongs to the quality of service scenario having 6.7 x 10-2 queuing delay in millisecond and the lower pane belongs to normal scenario without quality of service having a value of 1.52 millisecond queuing delay.

Queuing Delay (milliseconds) comparison

Ftp Down Response Time

Figure 5.7 the difference between the two scenarios with respect to download response time. Ftp download response time without QoS is 194 milliseconds and with QoS is 181 milliseconds. As mentioned in chapter 4 some changes were made. The Ftp server was changed from MK_DEPT network to PROD_DEPT network as more Ftp clients are present in that network.

Figure 5.8 show the impact of Ftp server changes, this change is compared with the QoS scenario to check whether the change gives some good significance or not. From the reading it does really give a good result as the Ftp download response comes up to 184 milliseconds as compared to 194 milliseconds.



Ftp Download response time (milliseconds) comparison



Ftp server location change



Emails download response time comparison

Ftp Server Load

Ftp server load without QoS

Ftp Server with QoS

After relocating the position Ftp server the value in traffic received bytes / second parameter also gave an impact. Without QoS the utilisation of server is 39% and number of bytes received is 632517. With QoS its utilisation is 53% and that's 1010898 bytes received by the server.

Utilisation of server after relocating is 36% which is lesser than 39% and the number of bytes is 613140.


Implementation of quality of service parameter and relocating the Ftp server put a significant impact on the behaviour of the network traffic and improved its efficiency in respect of download response timing, queuing delay, and link utilisation. It has helped to improve the over role network and make it scalable for the heavy traffic such as Video and Audio.


This scenario is the same as (MK_PROD_IPCLOUD) but it has no implementation of QoS. Without QoS this scenario compares with the IP scenario with QoS and at this stage it is the right time to check the performance of two giant wide area networks (WAN) even though one lacks QoS.

Central Link Utilisation (%)

The IP and ATM utilisation are 8.64% and 6.39 respectively. See figure 5.11. As a matter of fact ATM is fast than IP because of its technology and architecture.

IP central link utilisation

ATM central link utilisation

Queuing Delay

A very fine result is found in Queuing delay. In IP the value is 6.7 x 10-2 while in ATM the value is 8.1 x 10-3. See figure 5.12 (a and b) A straight line depicts a completely delegable delay.

Ftp Download Response Time

181 milliseconds and 175 milliseconds download response time is found in between IP and ATM network link. See figure 5.13.



Ftp Download response time (IP vs. ATM)


After having a critical comparison of the above two scenarios, it has been confirmed that ATM links are efficient and much faster than IP links. Data traffic having the same parameters when executed over IP and ATM networks, ATM gives different values and better functionality as compared to IP network.


These are the heaviest scenarios and most important as well where the true performance has been checked by deploying audio and video traffic.


a) Video packet end to end delay

IP video packet end to end delay = 7.59 ms

ATM video packet end to end delay = 5.76 ms



IP Vs ATM video packet end to end delay

b) Video packet delay variation



Video packet delay variation

IP video packet delay variation = 1.1 x 10-3

ATM video packet delay variation = 7.7 x 10-4


a) Voice packet end to end delay



Voice packet end to end delay

b) Voice Jitter


MK_PROD_ATM_AUDIO_VIDEO (bottom graph)

Voice jitter

c) Voice packet delay variation



Voice packet delay variation

iii) Central Link Utilisation

Central link utilisation ATM

Central link utilisation IP




Quality of service is used to boost up the network traffic in many aspects like; throughput, delay, and congestion control. This appendix illustrates how QoS was implemented in the simulation.

i) Open the object palette.

ii) Drag the object of QoS profile in the work space for implementation.

iii) Open the profile attribute window by right clicking and selecting "edit attribute", then select the WFQ (Weighted fair queue) profile to implement it. A popup window in the figure below will appear.

Quality of service profile attributes

iv) Open the WFQ tables and select "Queues Configuration" of ToS (Type of Service) profile, then give the weight of the network traffic according to the priority, for example: Higher the weight higher the priority.

Type of service weight values

Queue configuration table

iv) To define services to their respective weight values click on "Classification scheme" column of each weight. Shown in the figure below.

Assigning Type of services

v) Close the QoS profile object. Now the implementation process starts at this point.

Router_1 is selected in the existing network where QoS profile is assigned on its link where the data is transmitted and received.

vi) Open the attribute window of the router and select the "IP QoS Parameter" under IP.

vii) Select the "interface information" to implement the quality of service over the

Interfaces. See figure A. 5.

viii) After that, select the "QoS scheme" column for the selection of the queue type and queuing policy, is going to implement on that particular interface in this case select "WFQ Queue" type. See figure A. 6

IP QoS parameter table

Interface selection

Selecting Queue type


  • [Accessed July 16 th 2009 at 14:41]
  • [Accessed July 18, 2009 at 20:34]
  • Computer Networks Fourth Edition Andrew S. Tanenbaum page 291].
  • [ Accessed August 4, 2009 at 19:19]