ABSTRACT
This study showed that the network performance of a flat switch network in
the main library complex of Ambrose Alli University (AAU), Ekpoma can
greatly be improved by logical segmentation. A survey of the flat switch
network (that span across three departments with one broadcast domain) of
the library complex was carried out to ascertain the physical and logical
topology of the network and the number of hosts and network devices
available. The kind of traffic transmitted over the network was also
considered. Riverbed Modeler Academic Edition was used to simulate two
replicas of the library network. One of the simulated networks was logically
segmented by implementing Virtual Local Area Network (VLAN). Statistics
like traffic dropped, traffic forwarded, traffic received, broadcast traffic
dropped and traffic sent in bits/sec or packets/sec were collected from both
simulations and the results were analyzed and compared. The results from
the simulations showed that the application of VLAN immensely enhanced
the network performance by about 75%(depending on the size of the
network) because the logical segmentation increased the number of
broadcast domain while reducing each of the broadcast size. This further
implied that poor network design and large broadcast domain in a network,
gravely affects the performance of a network.
TABLE OF CONTENTS
TITLE PAGE i
CERTIFICATION ii
DEDICATION iii
ACKNOWLEDGEMENT iv
ABSTRACT x
1.0 INTRODUCTION 1
1.1 Background to the Study 1
1.2 Justification of the Study 2
1.3 Objectives of the Study 4
1.4 Research Methods 4
2.0 LITERATURE REVIEW 6
2.1 Review of related work/studies 6
2.2 Local Area Network Switching 7
2.3 Virtual LAN 10
2.3.1 Overview of Virtual Local Area Network 10
2.3.2 Identifying VLANs 14
2.3.3 VLAN Identification Methods 16
2.3.4 Routing between VLANs 17
2.3.5 Advantages of using VLANs 18
3.0 METHODOLOGY 19
3.1 Modelling and Simulation 19
3.2 Advantages of Simulation over live testing 20
3.3 Measurement Site/Environment 21
3.4 Statistics used to investigate Network Performance 26
3.5 Simulation/Measurement Software 28
3.5.1 Platform definition and description 28
3.5.2 Riverbed Modeler workflow 29
3.5.3 Building a Topology 31
3.5.4 Choosing appropriate statistics 34
3.5.5 Running a simulation and viewing results 36
3.6 Hosts and Network devices‘ Parameters 38
3.7 Simulation of the Library Complex Topology 40
3.8 Simulation of an Improved Switch Network 42
4.0 RESULTS AND DISCUSSION 44
4.1 Results 44
4.2 Discussion 51
5.0 CONTRIBUTION TO KNOWLEDGE, CONCLUSION AND
RECOMMENDATION 53
5.1 Contributions to Knowledge 53
5.2 Conclusion 53
5.3 Recommendation 54
LIST OF RESEARCH PUBLICATION 55
REFERENCES 56
CHAPTER ONE
INTRODUCTION
1.1 Background to the Study
As computer network grows in a campus environment, management of
dozens, hundreds, or even thousands of computers becomes increasingly
difficult. On a large, flat (Single Broadcast Domain), switched network,
performance suffers and security concerns increases (Deb, 2005). By
default, switches forward broadcasts from one network to another. Consider
a situation where you have 8 different devices connected to a switch. Since
switches always forward broadcasts to all clients on the network even
though the message is intended for one client, the network traffic is flooded.
As the number of devices connected to the switch increases, the amount of
bandwidth used by unnecessary broadcasts increases (Chris, 2004). One way
to structure a growing network is to divide it into segments called Virtual
Local Area Networks, or VLANs. Computers of users who work together
(workgroups) can be grouped into the same VLANs, even though they are
not located in close physical proximity. Often organizations create separate
VLANs for different departments or divisions. The VLAN serves as a
security boundary and improves performance by isolating broadcast and
multicast traffic (Deb, 2005). If users are grouped into multiple departments,
there is no need for users not in their department to receive these broadcasts.
Preventing those broadcasts from reaching unnecessary users can save a
large amount of bandwidth. Perhaps for security purposes, each department
needs to be kept separate on the switch, such that the clients systems in each
department are not reachable from ports in other departments (Deb, 2005).
VLANs allow this kind of logical grouping. Layer 2 switches will forward
frames between ports in the same VLAN, but will not do so between ports
not in the same VLAN (Deb, 2005).
By putting users whose jobs require constant use of the network to interact
with a particular server on the same VLAN segment with the servers to
which needs access, will prevent other users on separate VLANs from
having their network transactions with other servers impacted by the highusage
users. Their network performance won’t suffer, and the performance
of the high-usage users will also improve. This important and useful
arrangement is absent in AAU network, hence the need for this work.
1.2 Justification of the Study
Normally, a local area network (LAN) acts as a broadcast domain. That is,
all devices on the LAN will receive broadcast messages from all other
devices on that LAN. The communications within the LAN go through
devices such as hubs, switches and bridges. This will increase the amount of
traffic going to all devices and thus decreases network performance (Deb,
2005). Decrease in network performance could also be as a result of
improper allocation of users to different broadcast domains.
Poor network performance as a result of putting heavy and light users in one
broadcast domain could become major concern if not properly handled. All
users in a common broadcast domain shares network bandwidth. Network
performance decreases with increase in users accessing network resources
simultaneously. The decrease in performance is as a result of broadcast to
the entire network, and network congestion due to high network traffic. This
could also be a huge security concern (Todd, 2014).
A brief analysis of the network of AAU main library complex shows that the
entire network is one broadcast domain. The network covers SocketWorks,
Main Library and the ICT centre which has over hundred hosts on the
network. This shows that network performance and security is a concern.
The justification of this research stems from the fact that in most cases, poor
network performance is experienced during daily peak periods. The daily
peak periods includes office hours when most users are trying to access
resources over the network. This research intends to address the problem of
poor network performance as a result of large number of host in a single
broadcast domain. This research will therefore try to proffer solutions to the
problem of poor network performance especially during peak hours. The
research will be an eye-opener to the effects of poor network performance
and how it can be addressed.
1.3 Objectives of the Study
The overall aim of this research is to simulate an improved switch
network, using AAU main library complex as a case study.
The specific objectives of this research are to;
i. simulate the network topology of AAU main library complex,
ii. simulate an improved network topology for AAU main library complex,
iii. choose appropriate statistics that will be collected to generate results for
investigating the two simulated networks,
iv. analyze and compare results of the two simulated networks.
1.4 Research Methods
The research method employed includes;
i. Carrying out analysis of AAU main library complex. The network would
be inspected to analyze the physical and logical network topology, the
types and average number of network devices and hosts available on the
network. The type of network resources accessed by hosts will also be
put into consideration. Based on the network analysis done on the
network, a replica of the network topology will be simulated with
Riverbed Modeler Academic Edition.
ii. An improved network topology for AAU main library complex will be
simulated with Riverbed Modeler Academic Edition. The network
performance will be improved by implementing Virtual Local Area
Network (VLAN).
iii. Appropriate statistics like traffic dropped, traffic forwarded, traffic
received, and broadcast traffic dropped will be used to analyze and
investigate the experiment to generate useful results.
iv. The generated results of both simulated networks will be compared and
then the findings will be presented.
Do you need help? Talk to us right now: (+234) 08060082010, 08107932631 (Call/WhatsApp). Email: [email protected].
IF YOU CAN'T FIND YOUR TOPIC, CLICK HERE TO HIRE A WRITER»
