Download Thesis for the Master of Science degree by

Simulation and Analysis of Wireless Mesh Network in
Smart Grid – Advanced Metering Infrastructure
by
Philip Huu Huynh
B.A., The University of Economics, Vietnam, 1996.
A thesis submitted to the Faculty of Graduate School of the
University of Colorado at Colorado Springs
in partial fulfillment of the
requirements for the degree of
Master of Science
Department of Computer Science
2011
ii
Thesis for the Master of Science degree by
Philip Huu Huynh
has been approved for the
Department of Computer Science
by
_______________________________________________________
Advisor: Dr. C. Edward Chow
_______________________________________________________
Dr. Jugal K. Kalita
_______________________________________________________
Dr. Rory Lewis
_____________________
Date
iii
Simulation and Analysis of Wireless Mesh Network in
Smart Grid – Advanced Metering Infrastructure
by
Philip Huu Huynh
Master of Science, Computer Science
Thesis directed by Associate Dean Professor C. Edward Chow
Department of Computer Science
Abstract
In this thesis the use of Wireless Mesh Network (WMN) technologies as the Advanced
Metering Infrastructure (AMI) for the collecting of meter data in real-time was proposed
and analyzed. A Google Maps mashup was developed to display the locations of the
meters and light poles, which can be selected for mounting the WiMAX or Wi-Fi
network devices. A NS-3 simulator was developed to simulate the network traffic of
meter data collection over the WMN and to allow us evaluating different topologies and
to see if their capacities are adequate to report all meter values to the data center within
one second. A JavaScript program was developed to analyze the meter density within
different service areas of Colorado Springs Utilities. The information can be used for
antennae placement planning.
We proposed a hybrid WiMAX and Wi-Fi mesh network to address the cost and
efficiency issues. In local service area, we can use lower cost Wi-Fi mesh network to
iv
connect smart meters to the WiMAX Base Stations. From those WiMAX Base Stations,
we utilize the long distance links provided by WiMAX point-to-point connection mode to
collect meters from local service areas to a handful of take out points. From take-out
points, high speed optical fiber connections are to be used transport the meter data to the
data center. We evaluate the design trade offs of these WMN choices through the NS-3
simulations. The simulation results show that with this type of WMN (6 WiMAX Base
Stations and 540 Wi-Fi Access Points), it is feasible to collect meter data from 150,000
smart meters within one second.
The Smart Grid Wireless Infrastructure Planning (SG-WIP) Google Maps mashup tool
can be integrated with the simulator and allow the planner to interactively adjust the
planning of wireless network devices such as WiMAX Base Stations, or Wi-Fi Access
Points on the light poles in certain service areas.
v
Acknowledgments
This thesis could not have been accomplished without the assistance of many people
whose contributions I gratefully acknowledge.
First, I am heartily thankful to my advisor, Dr. C. Edward Chow, whose encouragement,
guidance and support from the initial to the final level enabled me to develop an
understanding of the subject. I would like to thank Dr. Jugal K. Kalita and Dr. Rory
Lewis, for their support during the thesis development. The value and contribution of this
thesis has been increased very much by their helps. Many thanks to my friends, the NS-3
community, the Google Maps development community, the Professors and Staff in the
Computer Science department for helping me finish this thesis.
I would like to thank the Management of Colorado Springs Utilities, for their support to
improve the possibility of practical application of this thesis.
Finally, I would like to show my gratitude to my parents for bringing me here a
wonderful place for learning and development.
vi
Table of Contents
Chapter 1 Introduction .........................................................................................................1
1.1 Thesis Statement .......................................................................................................2
1.2 Thesis Goals ..............................................................................................................2
1.3 Thesis Contributions .................................................................................................3
Chapter 2 Background and Related Works .........................................................................4
2.1 Introduction to Smart Grid - Advanced Metering Infrastructure ..............................4
2.1.1 Growing the need for the Smart Grid (SG) .........................................................4
2.1.2 Growing the need for the Smart Grid (SG) .........................................................6
2.2 Introduction to Wireless Mesh Network ...................................................................8
2.2.1 What is the Wireless Mesh Network (WMN)? ...................................................8
2.2.2 Wireless Mesh Network - The wireless infrastructure solution for AMI .........10
2.3 Related Works .........................................................................................................11
2.3.1 Colorado Springs Utility AMI Wireless Infrastructure ....................................11
2.3.2 SkyPilot Synchronous Mesh Network Solution ...............................................11
2.3.3 EkaNetTM Smart Network – Wireless Mesh Network Solution .......................13
2.3.4 Coverage and Capacity of A Wireless Mesh Network, by H.Huang, et al ......15
2.4 Research Opportunities ...........................................................................................16
2.5 Summary ..................................................................................................................17
Chapter 3 Problem and Solution .......................................................................................18
3.1 Problem Statement ..................................................................................................18
3.2 Approach .................................................................................................................18
3.2.1 Develop a Network Model for Communication Infrastructure ........................18
3.2.2 Simulate the AMI meter data reporting process ...............................................19
3.2.3 Analyze the Simulation Results ........................................................................20
3.3 Summary ..................................................................................................................20
vii
Chapter 4 Planning and Implementing the Network Simulation .......................................22
4.1 Introduction to the Smart Grid Wireless Infrastructure Planning (SG-WIP) ...........22
4.2 Planning the Network Simulation ............................................................................26
4.3 Design the Network Simulation ...............................................................................27
4.3.1 Physical Network Model ...................................................................................27
4.3.1.1 The Hybrid WMN Architecture ..................................................................27
4.3.1.2 The WiMAX/Wi-Fi Network Infrastructure ...............................................29
4.3.1.3 An Overview of the NS-3 WiMAX Module ...............................................32
4.3.1.4 An Overview of the NS-3 Wi-Fi Module ....................................................36
4.3.2 The Application Model ......................................................................................38
4.3.2.1 The Client-Server Architecture ...................................................................38
4.3.2.2 The Generation of Meter Data Traffic ........................................................39
4.3.2.3 The NS-3 Server Application ......................................................................39
4.3.2.4 The NS-3 Client Application .......................................................................39
4.3.3 The WLAN Simulation Design .........................................................................39
4.3.3.1 Results Analysis and Conclusion ................................................................40
4.3.3.2 Results Analysis and Conclusion ................................................................40
4.3.3.3 Results Analysis and Conclusion ................................................................40
4.3.3.4 Results Analysis and Conclusion ................................................................40
4.3.4 The WNAN Simulation Design .........................................................................40
4.3.4.1 Topology Configuration ..............................................................................40
4.3.4.2 Application Configuration ...........................................................................41
4.3.4.3 Simulation Planning ....................................................................................41
4.3.4.4 Results Analysis and Conclusion ................................................................41
4.3.5 The WMAN Simulation Design ........................................................................42
4.3.5.1 Topology Configuration ..............................................................................42
4.3.5.2 Application Configuration ...........................................................................42
4.3.5.3 Simulation Planning ....................................................................................42
4.3.5.4 Results Analysis and Conclusion ................................................................42
4.3.6 The WAN Simulation Design............................................................................43
4.3.6.1 Topology Configuration ..............................................................................43
viii
4.3.6.2 Application Configuration ...........................................................................43
4.3.6.3 Simulation Planning ....................................................................................43
4.3.6.4 Results Analysis and Conclusion ................................................................43
4.4 Implementing the Network Simulation ....................................................................44
4.4.1 The WLAN Simulation .....................................................................................44
4.4.1.1 The NS-3 Script ...........................................................................................44
4.4.1.2 The Linux Shell Script ................................................................................45
4.4.2 The WNAN Simulation .....................................................................................45
4.4.2.1 The NS-3 Script ...........................................................................................45
4.4.2.2 The Linux Shell Script ................................................................................46
4.4.3 The WMAN Simulation ....................................................................................47
4.4.3.1 The NS-3 Script ...........................................................................................47
4.4.3.2 The Linux Shell Script ................................................................................48
4.4.4 The WAN Simulation ........................................................................................48
4.4.4.1 The NS-3 Script ...........................................................................................48
4.4.4.2 The Linux Shell Script ................................................................................49
Chapter 5 Simulation Results and Analysis .......................................................................50
5.1 The Simulation Experiments ....................................................................................50
5.1.1 The WLAN Simulation Experiment ..................................................................50
5.1.2 The WNAN Simulation Experiment .................................................................51
5.1.3 The WMAN Simulation Experiment .................................................................51
5.1.4 The WAN Simulation Experiment ....................................................................53
5.2 Simulation Results Data Collection .........................................................................53
5.3 The Simulation Results ............................................................................................54
5.3.1 The WLAN Simulation Results
....................................................................54
5.3.1.1 Experiment 1: WLAN topology with 50 SMs.............................................54
5.3.1.2 Summary of the WLAN Simulation Experiments ......................................55
5.3.2 The WNAN Simulation Results
....................................................................55
5.3.2.1 Experiment 1: The WNAN topology with 9 APs........................................56
5.3.2.2 Summary of the WNAN Simulation Experiments ......................................59
5.3.3 The WMAN Simulation Results........................................................................59
ix
5.3.3.1 Experiment 1: The WMAN topology with 10 GWs ...................................60
5.3.3.2 Summary of the WMAN Simulation Experiments .....................................61
5.3.3.3 Summary of the Experiments for WMAN Improved Design ....................61
5.3.4 The WAN Simulation Results ...........................................................................62
5.3.4.1 Experiment 1: The WAN Star Topology With 3 BSs ................................62
5.3.4.2 Summary of the Experiments for the WAN Improved Design ...................64
5.4 Simulation Results Analysis ....................................................................................65
5.4.1 The WLAN Simulation Results Analysis ..........................................................65
5.4.2 The WNAN Simulation Results Analysis .........................................................67
5.4.3 The WMAN Simulation Results Analysis.........................................................69
5.4.4 The WAN Simulation Results Analysis ............................................................72
Chapter 6 Lessons Learned ................................................................................................77
6.1 The Development of SG-WIP Planning Tool ..........................................................77
6.2 The Development of Smart Grid Simulation Model ................................................78
6.3 The Simulation Process in NS-3 ..............................................................................79
6.3.1 The Initialization Phase of Wireless Networks .................................................79
6.3.2 The Bugs in NS-3 Module Code .......................................................................79
Chapter 7 Simulation Limitations and Future Work..........................................................81
7.1 Display Simulation Results on the SG-WIP Planning Tool .....................................81
7.2 Alternative Method for Network Traffic Application Simulation ...........................82
7.3 Improve the Antennae Placement Algorithm ..........................................................82
7.4 Store the Real-time Meter Data in the Database Management System (DBMS) ....83
7.5 Evaluate the Performance of the Network Model with the AMI real-time..............83
Chapter 8 Conclusion .........................................................................................................84
Bibliography ......................................................................................................................86
Appendix A SG-WIP User Manual ...................................................................................90
A.1 Installation ...............................................................................................................90
A.1.1 System Requirements .......................................................................................90
A.1.2 The Database ....................................................................................................90
A.1.3 The SG-WIP Application .................................................................................90
A.2 The GUI Operations ................................................................................................92
x
A.2.1 Opening the Home page ...................................................................................92
A.2.2 Generating the WAN Topology........................................................................93
A.2.3 Generating the MAN Topology ........................................................................94
A.2.4 Generating the NAN Topology
....................................................................95
A.2.5 Generating the LAN Topology .........................................................................96
A.2.6 Changing the Antennae Location of Network Devices ....................................97
Appendix B SG-SIM Smart Grid Simulator Running Examples ......................................98
B.1 The LAN Simulation Examples ..............................................................................98
B.2 The NAN Simulation Examples ............................................................................100
B.3 The MAN Simulation Examples ...........................................................................104
B.4 The WAN Simulation Examples ...........................................................................107
xi
List of Tables
Table 4.1 - The ranges of housing unit density of the LAN, NAN, MAN topologies in
Colorado Springs. ..............................................................................................................24
Table 5.2 - The WLAN simulation results with 50 SMs. The statistical data are in one
second period. ....................................................................................................................55
Table 5.3 - The results of WLAN simulations with different number of SMs. The
statistical data in each row are the results of one simulation experiment .........................55
Table 5.4 - The simulation results for WNAN 3x3 grid topology with nine APs. The
statistical data are in one second period.
....................................................................58
Table 5.5 – The results of WNAN simulation. The statistical data in each row are the
results of one simulation experiment .................................................................................59
Table 5.6 – The simulation results for WMAN Point-to-multipoint topology with ten
GWs. The statistical data are in one second period. ..........................................................61
Table 5.7 – The WMAN simulation results. The statistical data in each row are the results
of one simulation experiment............................................................................................61
Table 5.8 - The simulation results of a WMAN topologies experiment that has 10
gateways. The number of meter data packets put in a sending UDP packet was changed
until the UDP packet is fragmented. ..................................................................................62
Table 5.9 – The simulation results of a WAN Point-to-point Star topology with three base
stations. The statistical data are in one second period .......................................................63
Table 5.10 – The WAN simulation results. The statistical data in each row are the results
of one simulation experiment............................................................................................64
Table 5.11 - The of simulation results of the WAN topologies experiments that have one
base station connected to the data center. The length of the optical fiber cable was
changed to evaluate the total processing delay at a BS .....................................................64
xii
List of Figures
Figure 2.1 - Smart grid overview (source: U.S. Department of Energy) .............................5
Figure 2.2 - Smart Grid Conceptual Framework (source: The National Institute of
Standards and Technology) ..................................................................................................6
Figure 2.3 - AMI overview (source: National Energy Technology Laboratory) .................7
Figure 2.4 – AMI enabled integrated utilities operations (source: California Energy
Commission Meter Scoping Study) ..................................................................................8
Figure 2.5 -IEEE 802.11s terms: A Mesh Portal (MPP) connects to the wired Internet, a
Mesh Point (MP) just forwards mesh traffic, and a Mesh Access Point (MAP)
additionally allows stations (STA) to associate with it ........................................................9
Figure 2.6 - A SkyExtender was installed on the street light pole (source: SkyPilot) .......12
Figure 2.7 - SkyPilot Mesh Network Architecture (source: SkyPilot) ..............................12
Figure 2.8 - EkaNet Smart Network Architecture for AMI ...............................................14
Figure 2.9 - Mesh cell architecture for the outdoor applications .......................................16
Figure 4.1 – The SG-WIP tool for planning AMI wireless infrastructure network in
Colorado Springs ...............................................................................................................23
Figure 4.2 - Using SG-WIP tool for planning the antennae position. The WiMAX/Wi-Fi
gateway was place on streetlight pole ................................................................................24
Figure 4.3 - The WLAN topology (100x100 square meters) has a high density of resident
housing units ......................................................................................................................25
Figure 4.4 - WiMAX as backhaul inter Wi-Fi mesh networks (source: Intel) ..................29
Figure 4.5 - Logical view of the WM Communication Network includes the first three
layers of the OSI model
............................................................................................30
Figure 4.6 - Physical model of the WM Communication Network. The network hierarchy
includes the Wi-Fi Mesh Routers, the WiMAX/Wi-Fi Gateways, and the WiMAX BS ..31
xiii
Figure 4.7 - NS-3 WiMAX protocol stack overview .........................................................36
Figure 4.8 - NS-3 Wi-Fi layer 2 stack overview ................................................................38
Figure 4.9 - Smart meters access the Meter data center through the Wireless mesh
communication network.....................................................................................................39
Figure 5.1 - The simulation results in every one second for the WLAN Infrastructure
mode topology with one AP and fifty SMs in the network. .............................................65
Figure 5.2 - The simulation results for the WLAN infrastructure mode topology. The
number of SMs is assigned from the one to one hundred in the experiments to evaluate
the changing of the total processing delay at the SMs. ......................................................66
Figure 5.3 - The simulation results in every one second for the WNAN 3x3 mesh
topology with nine APs and one GW. AP sent 50 packets in every one second to the GW.
............................................................................................................................................67
Figure 5.4 - The simulation results of the WNAN 3x3 grid topology with nine MRs and
one GW in the network ......................................................................................................68
Figure 5.5 - The simulation results of WMAN point-to-multipoint topology with one BS
and ten GWs in the network. GW sent 180 packets in every one second to the BS ..........69
Figure 5.6 - The simulation results for the WMAN topology. The number of GWs are
assigned from the one to ten in the experiments to evaluate the changing of total
processing delay at the GWs ..............................................................................................70
Figure 5.7 - Impact on the network performance by aggregating meter data ....................71
Figure 5.8 - The simulation results in every one second for the WAN star topology with
DC and three BSs in the network. BS sent 1,800 packets in every one second to the DC 73
Figure 5.9 - The simulation results for the WAN star topology from many experiments .74
Figure 5.10 - The simulation results for the WAN star topology with one DC and one
BS. The length of the optical fiber cable was assigned from the one to 100 kilometers ..75
Figure A.1 - The SG-WIP application was installed on the Apache web server ...............91
Figure A.2 - The home page of SG-WIP ...........................................................................92
Figure A.3 - The WAN topology has twelve WiMAX base stations ................................93
Figure A.4 - The WMAN topology has WiMAX BS, and WiMAX/Wi-Fi Gateways .....94
xiv
Figure A.5 - The WNAN topology has WiMAX/Wi-Fi gateway, and Mesh
routers/Access Points .........................................................................................................95
Figure A.6 - The WLAN topology includes Wi-Fi Access Point, and Smart meters ........96
Chapter 1
Introduction
Recently many utilities started to deploy smart grids for collecting meter data [DoE01,
NETL08]. The main reasons are to reduce the cost by not sending people to read the
meter data and by avoiding generating excess power through correct prediction of the
load profile using the aggregated meter values. To correctly predict the load profile and
perform load forecasting, utilities need to collect meter data in real time.
Utilities have hundreds of thousands of meters installed in their service areas, and want to
network these meters for metering collection. The wireless communication technologies
have been popularly deployed in the local areas and the metropolitan areas because of
their conveniences in the cost, network installation and maintenance. Taking advantages
of the wireless technologies, utilities can network their meters and the data center for data
communication. However, if the underlined wireless technologies do not provide enough
bandwidth, then the meter data cannot be delivered to the data center in time. The
WiMAX technology allows us networking the meters and the data center with the
broadband data transmission at long distance and higher bandwidth [IEEE16, ZR06].
2
Therefore, the wireless networking solution for real-time metering collection is feasible.
The important question is to design of the wireless infrastructure and their topology so
that it can be scaled up and meet the cost and real-time performance requirements, given
huge wireless meters to be served in
large areas. For example, The Wi-Fi mesh
technology can be employed as part of a hybrid wireless infrastructure with WiMAX and
Wi-Fi to allow the deployment at the reasonable low cost [AWW05, INTL04].
The wireless technologies such as WiMAX, and Wi-Fi are high performance, scalable,
and secured [AWW05]. Taking the advantages of these network technologies, utilities
can deploy the smart grid wireless communications infrastructure for the real-time
metering collection. The real-time meter data can save the operation costs and reduce the
electricity market price.
1.1 Thesis Statement
In this thesis we plan to address the following important question: Is the wireless mesh
network infrastructure applicable for the real-time meter data collection?
It is a
challenge facing the smart grid wireless infrastructure planners. We intend to conduct a
simulation analysis of the wireless communication infrastructure for the smart grid to
answer this question.
1.2 Thesis Goals
In this thesis, we propose to research and develop a wireless communication
infrastructure solution using the wireless mesh network technologies for the smart grid.
Tools and techniques will be developed for the planning and designing the wireless
infrastructure. The performance and scalability properties of the proposed wireless
3
infrastructure are evaluated. We focus on these network properties because they are the
important factors that affect the performance of the real-time meter data collection.
1.3 Thesis Contributions
This thesis will contribute to the smart grid research by investigating the wireless mesh
network that is employed as a communication infrastructure solution for the real-time
metering collection. The wireless infrastructure planning tools in this thesis will benefit
not only the researchers but also the utilities. The network infrastructure planners and
researchers can use the planning tools to conduct surveys about wireless network
topologies. Moreover, the infrastructure planners and designers can use the tools to refine
their network designs.
Another contribution of this thesis is to provide a communication network solution that is
low cost, and secured. It allows the utilities to have an alternative option for their AMI
wireless infrastructure. The proposed wireless infrastructure with the wireless mesh
technologies are cheap, far-reaching, and scalable.
Chapter 2
Background and Related
Works
2.1 Introduction to Smart Grid - Advanced Metering
Infrastructure
2.1.1 Growing the need for the Smart Grid (SG)
A smart grid [DoE01, DoE02, Wiki01] delivers electricity from suppliers to consumers
using two-way digital communication technology. Smart grid allows controlling
appliances at consumer’s homes to save energy, and reduce cost. The operation status of
the smart grids can be monitored in real time, so the smart grids are more reliable. Many
governments are promoting such a modernized electricity network as a way of addressing
energy independence, global warming and emergency resilience issues.
5
Figure 2.10 - Smart grid overview (source: U.S. Department of Energy)
Figure 2.1 shows an overview of the smart grid. Utilities can archive energy efficiency
and maintain the competitive of services by taking advantages of the smart grid and its
market benefits. The smart grid solutions that utilize the information technology for data
collection, monitoring and control, data analysis and information communication
infrastructure, will cost-effectively protect revenues today, while laying the foundations
for future services.
Figure 2.2 shows the conceptual framework of smart grid. The components include
Service Provider, Operations, Markets, Bulk Generation, Transmission, Distribution, and
Customer.
6
Figure 2.11 - Smart Grid Conceptual Framework (source: The National Institute of
Standards and Technology)
2.1.2 Advanced Metering Infrastructure (AMI)
Advanced Metering Infrastructure (AMI) [NETL08] that is as part of larger Smart Grid
initiatives, is implemented by government agencies and utilities to meet the above
challenges. Extending from the current Automatic Meter Reading (AMR) technology,
AMI provides two way meter communications, and allows commands to be sent toward
the home for multi purposes, including Time-of-Use pricing information, demandresponse actions, or remote service disconnects.
7
Figure 2.12 - AMI overview (source: National Energy Technology Laboratory)
The Department of Energy estimates that over 280 Giga-watts of new generating capacity
will be needed by 2025. It results in new power plants would be built in the future. The
energy industry is facing the critical issues such as the need for new plants, maintaining
overburdened infrastructure, coping with an aging workforce, complying with
regulations, and environmental concerns. For a long time, the energy industry has
rightfully focused on the supply side of this challenge. But now, the demand side of the
equation can be significantly impacted by the existing of the wireless mesh networking
[Kri01].
8
Figure 2.13 - AMI enabled integrated utilities operations (source: California Energy
Commission Meter Scoping Study)
Wireless mesh networking can use as the backbone of the AMI solutions to enable twoway intelligent networked communications with smart meters. With the AMI, the value
added services such as demand response and demand side management would be
enabled, besides meter reading. AMI solutions allows interoperable networks and
systems across the entire power structure aid in the management and control of energy
consumption, improve operations management, conserve the environment, and adhere to
evolving regulations [NETL08].
2.2 Introduction to Wireless Mesh Network
2.2.1 What is the Wireless Mesh Network (WMN)?
9
A WMN [IEEE11s, Moh01] is a communications network made up of radio nodes
organized in a mesh topology. Wireless mesh networks often consist of mesh clients,
mesh routers and gateways. The mesh clients are often laptops, cell phones and other
wireless devices while the mesh routers forward traffic to and from the gateways which
may but need not connect to the Internet.
A mesh network is reliable and offers
redundancy. When one node can no longer operate, the rest of the nodes can still
communicate with each other, directly or through one or more intermediate nodes.
Wireless mesh networks can be implemented with various wireless technology including
IEEE 802.11, IEEE 802.15, and IEEE 802.16 [IEEE11, IEEE15, IEEE16] , cellular
technologies or combinations of more than one type.
Figure 2.14 - IEEE 802.11s terms: A Mesh Portal (MPP) connects to the wired
Internet, a Mesh Point (MP) just forwards mesh traffic, and a Mesh Access Point
(MAP) additionally allows stations (STA) to associate with it
10
2.2.2 Wireless Mesh Network - The wireless infrastructure solution
for AMI
When the WMN technology is applied in AMI solutions, it can bring new components to
the electrical grids, such as self-managing and self-healing mesh networking, intelligent
meters, and bridging to Home Area Networks (HAN) [NETL08] for connectivity with
energy consuming appliances. The real time communication between the smart meters
and the utility’s data center provides detailed usage data while also receives and display
Time-of-Use (TOU) pricing information, and offers other on-demand abilities such as
remote connect or disconnect, unrestricted monitoring and control, etc. Customers are
able to access the usage data for tailoring consumption and minimizing energy expenses
while helping balance overall network demand.
When WMNs are used in the AMI, they can provide the following features [LLT03,
Kri01]:

Low cost of management and maintenance - WMNs are self-organizing and
require no manual address/route/channel assignments. It is simple to manage
thousands or millions of devices resulting in the lowest total cost of ownership.

Increased reliability – The WMN routing mechanisms provide the redundant paths
between the sender and receiver of the wireless connection. Communication
reliability is significantly increased because of the eliminations of single point
failures and potential bottleneck links. Network robustness against potential
problems, e.g., node failures and path failures due to RF interferences or obstacles,
can also be ensured by the existence of multiple possible alternative routes.
11

Scalability, flexibility and lower costs - WMNs are self-organizing and allow true
scalability. Nodes and Gateways are easily added at a very low cost with:
o No limitation on number of hops
o No network address configuration
o No managed hierarchical architecture
o No hard limitation on number of Nodes per Gateway

Robust security – WMNs have the security standards that allows all
communications in AMI are protected by mutual device authentication and derived
per-session keys using high bit rate AES encryption. This hardened security
approach allows for authentication as well as confidentiality and integrity
protection in each communication exchange between every pair of network
devices – Smart meters, Relays, or Wireless Gateways.
2.3 Related Works
2.3.1 Colorado Springs Utility AMI Wireless Infrastructure
The Colorado Springs Utility AMI wireless network infrastructure used the Point-toMulti Point topology where 902-928MHz concentrators are used to collect meter data in
a neighbor area mounting on the light pole, eight take-out points are used to poll and
collect meter data from hundreds of concentrators. Telecommunication links and fiber
connections are used to connect take-out points to the data center. The current meter data
reading interval is five minutes for electric, fifteen minutes for gas and water meters
[Chow09].
12
2.3.2 SkyPilot Synchronous Mesh Network Solution
MetroFi has deployed a mesh network in the Silicon Valley – California [Met01]. The
installed wireless mesh metropolitan area network can provide the Internet access service
to the resident user in a geographical area that covers about 250,000 households. The
SkyPilot’s Synchronous Mesh Network solution was employed to build this mesh
network. The SkyPilot’s mesh network solution combines standard-based Wi-Fi access
with a high performance wireless mesh backhaul network using SkyGate nodes to
interface with the Internet, SkyExtender DualBand nodes that integrate Wi-Fi access and
mesh backhaul [Sky01].
Figure 2.15 - A
SkyExtender was
installed on the street
Figure 2.16 - SkyPilot Mesh Network Architecture (source: SkyPilot) light pole (source:
SkyPilot)
The mesh network MetroFi is a success deployment of the WMN onto the large
geographical area. However, in the AMI meter data collection application, there is a
different in the network traffic pattern compared to the regular Web applications. The
13
Web applications usually need the connection with low bandwidth uplink, and high
bandwidth downlink. In contrast, the AMI meter data reporting process requires the high
bandwidth uplink connections to send the data from smart meters to data center.
We can make an assumption that we intend to use the MetroFi network for the AMI
communication infrastructure solution. Then, there is a research opportunity about the
MetroFi or WMN performance – Whether the WMN is suitable for the AMI network
infrastructure, especially for the real-time meter data reporting application?
2.3.3 EkaNet™ Smart Network - Wireless mesh network
solution for Smart Grid/AMI
The wireless mesh infrastructure EkaNet [Eka01] includes the smart meter nodes, ranger
extension nodes, and gateway nodes. The smart meter nodes are networking together to
form the wireless mesh network. It means that, the communication between a smart meter
node and the gateway will replay on a number of other smart meters. The range extension
nodes are used to help connect the out of coverage nodes. The gateway nodes provide the
interface to the Internet network.
Figure 2.17 - EkaNet Smart Network Architecture for AMI
14
The EkaNet has been deployed in some world wide areas such as Guayaquil - Ecuador,
St. Petersburg – Russia. In Guayaquil – Ecuador, the utility has installed 3,614 wireless
meter nodes, 314 repeaters, and 47 gateways. Meter data collection interval period is
fifteen minutes.
The Advantages
Wireless mesh network provides a low cost, easy deployment and management,
scalable, flexible
The Disadvantages
The number of relay hops will increases with the increasing number of the smart
meters. As a result, the network performance will go down fast, especially in the
service place where the smart meter density is high, i.e. hundreds of smart meters
in the area 100 meters by 100 meters. In the high resident density area, the mesh
topology will not be a good choice for the network performance goal. Instead, the
point-to-multipoint topology such as Wi-Fi infrastructure mode would provide a
better network performance.
2.3.4 Research Paper “Wireless Mesh Networks: A
Survey” [AWW05]
The authors conducted a survey about the current research and development of WMN
technologies (2005). The author presented many open research issues needed to be solved
such as scalability, self-organization and self-configuration, security, network integration.
15
The critical factors influencing protocol design were discussed for improvement
objectives.
2.3.5 Research Paper “The Nominal Capacity of Wireless
Mesh Networks” [JS03]
This article tackles the problem of determining the exact capacity of a WMN. The
concept of bottleneck collision domain was introduced to enable this calculation. The
authors shown that for WMNs the throughput of each node decreases as O(1/n), where n
is total number of nodes in the network. Moreover, for a given topology and the set of
active nodes, the upper bounds on the throughput of any node can be exactly calculated.
2.3.6 Research Paper “Capacity of Grid-Oriented Wireless
Mesh Networks” [ANMK08]
The author presented an analytical framework for determining the nominal capacity of
multi-radio multi-channel Wireless Mesh Network (WMN). The concept of collision
domain was derived to calculate an upper bound on the capacity available to ingress
nodes that generate traffic towards the node that acts as the gateway to the outside world.
As the research conclusion, the effects of WMN design parameters such as network
topology, network size, routing methods, channel assignment schemes etc. are interlinked
and a judicious selection is essential to maximize capacity.
2.3.7 Research Paper “Architecture and Algorithms for an
IEEE 802.11-based Multi-channel Wireless Mesh
Network” [RC05]
16
The author proposed a novel multi-channel WMN architecture that effectively addresses
the bandwidth problem by fully exploiting non-overlapped radio channels that the IEEE
802.11 standards make available. Two fundamental design issues in the proposed multichannel WMN architecture are solved. The first one is about the determining which of
the available non-overlapped radio channels should be assigned to each 802.11 interface
in the WMN. The second issue is how packets should be routed through a multi-channel
WMN.
2.3.8 Research Paper “Multi-Channel Mesh Networks:
Challenges and Protocols” [KSCV06]
The authors considered the use of multi-channel to improve the throughput of Wireless
Mesh Network (WMN). The main challenges were highlighted and two link-layer
protocols were presented for utilizing multiple channels. A new abstraction layer that
simplifies the implementation of new multi-channel protocols in existing operating
systems was presented.
2.3.9 Research Paper “Coverage and capacity of a wireless
mesh network” [HWC05]
The authors proposed a scalable multi-channel ring-based WMN architecture and
developed an analytical framework to evaluate the capacity and coverage of such a
network. An optimization approach to maximize capacity and coverage for the
considered WMN has been also presented.
2.3.10 Research Paper “The IEEE 802.11s Extended
Service Set Mesh Networking Standard” [CK08]
17
The author presented how the developing IEEE 802.11s ESS Mesh Networking Standard
draft addresses technical challenges of the pervasive development of wireless mesh
networks (WMNs), the efficient allocation of mesh resources (routing and MAC layers),
the protection of network resources (security and power savings), and the elimination of
spatial bias (congestion control). Many examples from existing two-tier developments,
simulations, and analytical models are used to motivate these enhancements within the
standard.
2.3.11 Research Paper “An Improved IEEE 802.16
WiMAX Module for the ns-3 Simulator” [IPGT10]
The authors presented the new features and enhancements that were integrated within the
ns-3 WiMAX module. The new design of the physical layer has improved the simulation
time by several magnitude orders while still providing a realistic implementation of the
standard. The IP classifier has enabled the simulation of an unlimited number of service
flows per subscriber station, while the proposed schedulers improve the management of
the QoS requirements for the different service flows. These proposed features can make
easier and more realistic the evaluation and design of WiMAX systems.
2.4 Research Opportunities
This thesis discusses about the evaluation of performance of the WMN when it is
employed as the wireless infrastructure solution for the AMI real-time metering data
collection application.
The related work has proved that the WMN can be used in the networking solutions that
require the deployment onto large geographical areas, such as the AMI communication
18
infrastructure. However, the WMN infrastructure needs a high bandwidth for transmitting
the meter data from the smart meters to the data center in real-time. The conducted
researches have shown that the WMN network bandwidth is affected many network
design parameters such as network the topology, network size, routing protocols, multichannel assignment scheme [ANMK08, RC05, KSCV06]. In particularly, the more hop
number is the WMN routing path, the less performance is the WMN [LLT03, ZR06,
HWC08].
2.5 Summary
In general, the wireless mesh network infrastructure can provides a cheap solution,
compared to the wire network, connecting the smart meters to the utility data center.
However, to answer the challenge question, whether the wireless mesh infrastructure is
suitable for the real-time meter data reporting process, this thesis will go into more detail
in the analysis of the performance property of the
wireless network infrastructure
solution. We will develop new tools and techniques to assist the planning and design
phase. We will also use the network simulation method to evaluate the network
performance.
Chapter 3
Problem and Solution
3.1 Problem Statement
The WMN contributes many advantages to the AMI Communication Network solution.
However, there are challenge questions in planning, designing and deployment of WMN.
•
Does the WMN meet the network performance requirements for real-time Meter
Data Collection?
•
What is the trade-off between the Performance and Scalability, for cost
optimization?
This thesis will answer these challenge questions by using simulation method.
3.2 Approach
3.2.1 Develop a Network Model for Communication
Network
20
For the network researching, planning and designing, we will develop and implement a
network topology planning application. The application can assist the network planning
and design phase, for example the planning of antennae placement of the wireless
devices, or research the network traffic based on the smart meters density in the service
areas.
For the simulation of AMI real-time meter data reporting application, we will develop a
hybrid WMN model for AMI wireless infrastructure solution. The hybrid WMN model
employs a network architecture that uses the wireless mesh technologies, and the pointto-multipoint technologies to network many thousands of wireless nodes together for the
network communication. The hybrid WMN model uses the WiMAX (IEEE 802.16d) and
Wi-Fi (IEEE 802.11 a/b) technologies [INTL04, HWC08].
We are interested in the hybrid architecture because it is high performance and scalable.
These properties are very important because AMI meter data reporting application will be
deployed in the large areas.
3.2.2 Simulate the AMI Meter Data Reporting process
Many network simulation experiments will be created. The WMN simulation process is
divided into the smaller network topologies simulation processes. We can create and
simulate the simulation experiments for Wireless LAN, Wireless NAN, Wireless MAN,
and WAN topologies.
We also develop a network traffic generator application that simulates the real time meter
data reporting process from the smart meters to the data center.
The network simulation process can be implemented on the simulation software NS-3
[NS3].
21
3.2.3 Analyze the Simulation Results
The network throughput (the number of messages received in one second) will be derived
from the simulation results to see whether the communication network can transport the
meter data from all of the meters to the data center in one second.
The WMN model is also investigated about the trade-off between the scalability and the
performance, and that can help the optimization in the network designing phase.
3.3 Summary
For the WMN infrastructure solution, there are some issues concerning about the AMI
meter data collection application. We will briefly discus the issues and the solutions.
Firstly, the application deployment is throughout a large geographical area, such as a city
or a metropolitan. So, there is the need for the installation of the large wireless networks
or so called wireless metropolitan area network (WMAN). This issue can be
accomplished by carefully planning the network topology and capacity.
In this thesis, we will introduce the wireless networking solutions using the modern
wireless network technologies such as WiMAX and Wi-Fi. We also discuss the planning
and designing of the WMAN infrastructure using these wireless networking technologies.
We will develop the tools and techniques to assist the planning and designing process.
Secondly, how we can evaluate the performance of such larger wireless network
infrastructure. In the scope of this thesis, for evaluating network performance
measurement, we will use the network simulation method to accomplish this issue.
22
We will develop a network model for our wireless mesh network infrastructure solution.
Then we will simulate the network model using the network simulation software NS-3.
The simulation results will be analyzed for the evaluation of the network performance.
Chapter 4
Planning, and Implementing
the Simulation
4.1 Introduction to the Smart Grid Wireless
Infrastructure Planning (SG-WIP) Tool
The SG-WIP is a Wireless Network Topology Planning Application. We have developed
this planning tool to assist the planning, and designing phase of the AMI wireless
network infrastructure. Figure 4.1 shows the GUI of SG-WIP.
The SG-WIP is a Google Maps mashup [Goog01, PSC06]. It can provide the information
about the geographical location of the network topologies, network devices, or the
residential housing units in the service areas of the utility.
24
Figure 4.10 - SG-WIP tool for planning AMI wireless infrastructure network in
Colorado Springs
In the network planning phase, we have conducted some researches that use the SGWIP
tool.

The research for antenna placement of the WiMAX/Wi-Fi networks has employed the
SG-WIP platform as a tool to extract information of the geographical network
topologies such as housing unit locations, or street light poles.
Figure 4.2 shows the planning antennae placement for the smart meters and the
WiMAX/Wi-Fi gateway on the Google Maps.
25
Figure 4.11 - Using SG-WIP tool for planning the antennae position. The
WiMAX/Wi-Fi gateway was place on streetlight pole.

The research about housing unit density of the designing wireless networks has also
used the SGWIP platform to gather the distribution of the housing units.
Table 4.1 shows the range of number housing units in the LAN, NAN, WAN
topologies. The dimensioning information is helpful for the designing of smart grid
network simulation. For example, Table 4.1 shows the number of housing units in the
LAN, NAN, MAN topologies for the conducted simulation.
Low Bound
(housing units)
LAN
NAN
MAN
High Bound
(housing units)
Simulation
0
51
50
0
1,054
950
0
40,501
27,000
Table 4.12 - The ranges of housing unit density of the LAN, NAN, MAN topologies
in Colorado Springs.
26
Figure 4.3 shows the WLAN topology size 100x100 square meters that has fifty
housing units.
Figure 4.12 - The WLAN topology (100x100 square meters) has a high density of
resident housing units.
The exported information about the network topologies from SG-WIP platform, as well
as the research results about the housing unit density, and the antenna locations can help
the AMI network infrastructure researchers and designers in the simulation and analysis
of the wireless network infrastructure of the AMI.
27
4.2 Planning the Network Simulation

The following network topologies will be simulated:
o Wireless Local Area Network (WLAN)
o Wireless Neighborhood Area Network (WNAN)
o Wireless Metropolitan Area Network (WMAN)
o Wide Area Network (WAN)

The main purpose is for evaluating the network throughput of the Hybrid
WiMAX/Wi-Fi Infrastructure that will be employed for the AMI meter reading
reporting application
o
o
Network topologies

WiMAX, Wi-Fi technologies

Grid Topology: with pre-defined distance between wireless nodes

Adequate bandwidth data link connection
Applications

Traffic pattern: Up-link data flows from the Smart Meter nodes to the
Utilities Data Center node

Each Smart Meter sends one meter reading message to the Data Center in
every second. The network throughput is calculated based on the number
of arrived messages in every one second at the Data center.
o
The network throughput is measured from many simulation experiments that
have the inputs as following:

Number of Smart Meter nodes
28


Number of Wireless Mesh Hops, and Access Points

Number of WiMAX/Wi-Fi Gateways

Number of WiMAX Base Stations
The transmission delay (Tx Delay) of a meter data message is designed to measure
the average delay of the transmission of a meter data message throughout the network
infrastructure.
4.3 Design the Network Simulation
4.3.1Physical Network Model
4.3.1.1 The Hybrid WMN Architecture
There are three types of WMNs: Flat WMN, Hierarchical WMN, and Hybrid WMN
[HWC08]. The brief description for these WMN categories is as following:
4.3.1.1.1 Flat Wireless Mesh Network
The flat WMN includes nodes that have roles as both client and router. The nodes can
perform the networking functionalities such as routing, network configuration, services,
and other applications. This architecture is similar to the Ad-hoc wireless network and it
is the simplest type among the three WMN architecture types. Its disadvantages are lack
of network scalability and high resource constraints.
4.3.1.1.2 Hierarchical Wireless Mesh Network
The hierarchical WMN has multiple tiers or levels. The client nodes form the lowest tier
in the hierarchy. The client nodes communicate together through the backbone network
formed by WMN routers. The WMN routers are the dedicated nodes for routing
29
functions. They are not source or destination of data traffic like the client nodes. In the
backbone network, there are some router nodes that may have external connections to the
other resources such as the Internet, and other servers in wired networks, and then such
nodes are called gateway nodes.
4.3.1.1.3 Hybrid Wireless Mesh Network
Hybrid WMN is a special case of the hierarchical WMN where the WMN utilizes other
wireless networks for communication. For example, the hierarchical WMN that has the
client and router nodes used the Wi-Fi technology, can employ the infrastructure-based
networks such as cellular, WiMAX, or satellite networks to connect to the Internet.
The hybrid WMNs can utilize multiple technologies for both WMN backbone and
backhaul. Since the growth of the WMNs depend heavily on the ability to work with
other existing wireless networking solutions, this architecture type is very important in
the future.
In the figure 4.4, the WiMAX has been use directly as part of Wi-Fi mesh network. The
WiMAX Subscriber Terminal put on the Wi-Fi Mesh Access Point. So the Wi-Fi
Networks automatically are more reliable in wider coverage area, and reduce cost of
connections that are caused by cable drawing in the gateway installation.
30
Figure 4.13 - WiMAX as backhaul inters Wi-Fi mesh networks (source: Intel)
4.3.1.2 The WiMAX/Wi-Fi Network Infrastructure
Basically, the WM Communication Network component provides the data transportation
services. The requests and responses from Meter Data Center component and Wi-Fi
Smart Meter component will be delivered by the using to the transportation services of
WM Communication Network component.
The WM Communication Network component has three layers of network services like
the first three layers of the OSI model [OSI]:
31
Figure 4.14 - Logical view of the WM Communication Network includes the first
three layers of the OSI model
The WM Communication Network is an integrated Wireless Mesh Network (WMN),
which uses Wi-Fi and WiMAX technologies [INTL04]. The WM Communication
Network has the WiMAX Base Station, the WiMAX/Wi-Fi Gateway, and Wi-Fi Dual
Band Mesh Routers.
The figure 4.6 shows the physical model of the wireless mesh communication network.
The WiMAX Base Stations are connected to the Meter Data Center through wired
network. The Wi-Fi mesh routers are at the bottom level of the network hierarchy and can
connect with the Wi-Fi smart meters. Wi-Fi smart meters connect to the meter data center
via the hybrid WiMAX/Wi-Fi Communication Network.
32
Figure 4.15 - Physical model of the WM Communication Network. The network
hierarchy includes the Wi-Fi Mesh Routers, the WiMAX/Wi-Fi Gateways, and the
WiMAX BS.
33
4.3.1.3 An Overview of the NS-3 WiMAX Module
The NS-3 WiMAX model attempts to provide an accurate MAC and PHY level
implementation of the IEEE 802.16 specification with the Point-to-multipoint (PMP)
mode and the Wireless MAN-OFDM PHY layer. The WiMAX model composed of three
layers:

The MAC Convergence Sub layer (MAC-CS)

The MAC Common Part Sub layer (MAC-CPS)

The Physical (PHY) layer
The MAC Convergence Sub layer (CS)
The MAC-CS in this module implements the Packet CS, designed to work with the
packet-based protocols at higher layers. The CS is responsible of receiving packet from
the higher layer and from peer stations, classifying packets to appropriate connections (or
service flows) and processing packets. It keeps a mapping of transport connections to
service flows. This enables the MAC CPS identifying the Quality of Service (QoS)
parameters associated to a transport connection and ensuring the QoS requirements.
The MAC Common Part Sub layer (MAC-CPS)
The MAC Common Part Sub layer (CPS) is the main sub layer of the IEEE 802.16 MAC
and performs the fundamental functions of the MAC. The module implements the PointMulti-Point (PMP) mode. In PMP mode BS is responsible of managing communication
among multiple SSs. The key functionalities of the MAC-CPS include framing and
addressing, generation of MAC management messages, SS initialization and registration,
service flow management, bandwidth management and scheduling services.
34

Framing and Management Messages
The module implements a frame as a fixed duration of time, i.e., frame boundaries are
defined with respect to time. Each frame is further subdivided into downlink (DL) and
uplink (UL) sub frames. The module implements the Time Division Duplex (TDD) mode
where DL and UL operate on same frequency but are separated in time. A number of DL
and UL bursts are then allocated in DL and UL sub frames, respectively. Since the
standard allows sending and receiving bursts of packets in a given DL or UL burst, the
unit of transmission at the MAC layer is a packet burst. The module implements a special
Packet Burst data structure for this purpose. A packet burst is essentially a list of packets.
In the case of DL, the sub frame is simulated by transmitting consecutive bursts
(instances Packet Burst). In case of UL, the sub frame is divided, with respect to time,
into a number of slots. The bursts transmitted by the SSs in these slots are then aligned to
slot boundaries. The frame is divided into integer number of symbols and Physical Slots
(PS) which helps in managing bandwidth more effectively. The number of symbols per
frame depends on the underlying implementation of the PHY layer. The size of a DL or
UL burst is specified in units of symbols.

Network Entry and Initialization
The network entry and initialization phase is basically divided into two sub-phases, (1)
Scanning and synchronization and (2) Initial ranging. The entire phase is performed by
the LinkManager component of SS and BS.

Connections and Addressing
All communication at the MAC layer is carried in terms of connections. The standard
defines a connection as a unidirectional mapping between the SS and BS's MAC entities
35
for the transmission of traffic. The standard defines two types of connections: the
Management Connections for transmitting control messages and the Transport
Connections for data transmission. Note that each connection maintains its own
transmission queue where packets to transmit on that connection are queued. The
ConnectionManager component of BS is responsible of creating and managing
connections for all SSs.

Scheduling Services
The module supports the four scheduling services defined by the IEEE 802.16-2004
standard:
Unsolicited Grant Service (UGS)
Real-Time Polling Services (rtPS)
Non Real-Time Polling Services (nrtPS)
Best Effort (BE)
These scheduling services behave differently with respect to how they request bandwidth
as well as how it is granted. Each service flow is associated to exactly one scheduling
service, and the QoS parameter set associated to a service flow actually defines the
scheduling service it belongs to. When a service flow is created the UplinkScheduler
calculates necessary parameters such as grant size and grant interval based on QoS
parameters associated to it.
WiMAX PHY Model
The Wireless MAN OFDM PHY specification is implemented. This specification is
designed for non-light-of-sight (NLOS) including fixed and mobile broadband wireless
access. The proposed model uses a 256 FFT processor, with 192 data subcarriers. It
36
supports all the seven modulation and coding schemes specified by Wireless MANOFDM. It is composed of two parts: the channel model and the physical model.

Channel model
When a physical device sends a packet (FEC Block) to the channel, the channel handles
the packet, and then for each physical device connected to it, it calculates the propagation
delay, the path loss according to a given propagation model and eventually forwards the
packet to the receiver device.

Physical model
The physical layer performs two main operations: (i) It receives a burst from a channel
and forwards it to the MAC layer, (ii) it receives a burst from the MAC layer and
transmits it on the channel.
Transmission Process: A burst is a set of WiMAX MAC PDUs. At the sending process, a
burst is converted into bit-streams and then split into smaller FEC blocks which are then
sent to the channel with a power equal P_tx.
Reception Process: The reception process includes the following operations:
1- Receive a FEC block from the channel. 2- Calculate the noise level. 3- Estimate the
signal to noise ratio (SNR) with the following formula. 4- Determine if a FEC block can
be correctly decoded. 5- Concatenate received FEC blocks to reconstruct the original
burst. 6- Forward the burst to the upper layer.
The below figure 4.3 shows an overview of the WiMAX sub-layers traversed for
transmitting and receiving a packet. More detailed information about the NS-3 WiMAX
model is presented in [NS3].
37
Figure 4.16 - NS-3 WiMAX protocol stack overview
4.3.1.4 An Overview of the NS-3 Wi-Fi Module
The NS-3 802.11 model provides an accurate MAC-level implementation of the 802.11
specification and the PHY-level model of the 802.11a and 802.11b specifications.
There are four levels that were implemented in the current implementation:

The PHY layer model

The so-called MAC low models

The so-called MAC high models

A set of Rate control algorithms used by the MAC low models
38
The PHY layer implements a single 802.11a model in the ns3::WifiPhy class, and
recently extended to cover 802.11b physical layers.
The MAC low layer is split in 3 components:

ns3::MacLow takes care of RTS/CTS/DATA/ACK transactions

ns3::DcfManager and ns3::DcfState implement the DCF functions

ns3::DcaTxop and ns3::EdcaTxopN handle the packet queue, packet
fragmentation, and packet retransmissions.
The MAC high models contain the implementations for three Wi-Fi topological elements
– Access Point (AP) implemented in ns3::ApWifiMac, non-AP Station (STA)
implemented in ns3::StaWifiMac, and STA in an Independent Basic Service Set (IBSS)
implemented in ns3::AdhocWifiMac.
Rate control Algorithms include:

ns3::ArfWifiManager

ns3::AarfWifiManager

ns3::IdealWifiManager

ns3::CrWifiManager

ns3::OnoeWifiManager

ns3::AmrrWifiManager

ns3::CaraWifiManager

ns3::AarfcdWifiManager
The below figure 4.4 shows the overview of the Wi-Fi L2 sub-layers traversed for
transmitting and receiving a packet. More detailed information about the NS-3 Wi-Fi
model is presented in [NS3].
39
Figure 4.17 - NS-3 Wi-Fi layer 2 stack overview
4.3.2 The Application Model
4.3.2.1 The Client-Server Architecture
The AMI metering data collection process includes three components that are Meter Data
Center, Wireless Mesh (WM) Communication Network, and Wi-Fi (WF) Smart Meter.
The Meter Data Center component accesses the WF Smart Meter’s reading via the WM
Communication Network as in the Figure 4.9.
40
Figure 4.18 - Smart meters access the Meter data center through the Wireless mesh
communication network
4.3.2.2 The Generation of Meter Data Traffic
Our current software simulates constant bit rate traffic. We allow users specifying the
starting time of packet streams. This allows for better network performance since the
packets from different nodes will not collide. It also helps debug the end to end
transmission and ensures that the network properly delivers the packets.
4.3.2.3 The NS-3 Server Application
An UDP protocol Server. It receives the meter messages.
4.3.2.4 The NS-3 Client Application
An UDP protocol Client. It sends the meter messages to the Server.
4.3.3 The WLAN Simulation Design
4.3.3.1 Topology Configuration

Standard: Wi-Fi IEEE 802.11b

Connection mode: Infrastructure

Smart Meter (SM) at random position within the coverage area of the
corresponding AP
41

The Wi-Fi AP has the coverage range of 100 meters

Number of SMs: [1 – 100]

Wi-Fi link capacity: 11Mbps
4.3.3.2 Application Configuration

Server application is installed on the AP.

Client application is installed on SM.

Each Client application will send one meter message with 20 bytes length
to the Server application by using the Internet protocol UDP.

The Client application’s Data-Rate property is set to 20 bytes x 8 bits =
160bps = 0.160kbps
4.3.3.3 Simulation Planning

Repeatedly running the simulation scenarios with the different number of
SMs

Output: the network throughput, Tx Delay
4.3.3.4 Results Analysis and Conclusion

Calculate the average network throughput, Tx delay

Conclusion: Do the AP receive all of the messages from the SMs in 1
second?
4.3.4 The WNAN Simulation Design
4.3.4.1 Topology Configuration

Standard: Wi-Fi IEEE 802.11a

Connection mode: Mesh
42

The Mesh Routers (MR) /Access Points (AP) are installed in the Grid
topology
o Distance between adjacent nodes (horizontal and vertical): 200
meters

Number of MRs/APs: [1 – 9]

Wi-Fi link capacity: 54Mbps
4.3.4.2 Application Configuration

Server application is installed on the Gateway (GW).

Client application is installed on APs.

Each Client application will send 100 messages, which have 20 bytes
length, to the Server application by using the Internet protocol UDP.

The Client application’s Data-Rate property is set to 100 x 20 bytes x 8
bits = 16000bps = 16kbps
4.3.4.3 Simulation Planning

Repeatedly running the simulation scenarios with the different number of
MRs and APs

Output: the network throughput, Tx delay
4.3.4.4 Results Analysis and Conclusion

Calculate the average network throughput, Tx delay

Conclusion: Do the GW receive all of the messages from the APs in 1
second?
43
4.3.5 The WMAN Simulation Design
4.3.5.1 Topology Configuration

Standard: WiMAX IEEE 802.16d

Connection mode: Point-To-Multipoint

The Subscribers (SS)/Gateways (GW) are installed in the grid topology.
Distance between adjacent nodes (horizontal and vertical): 1,000 meters

Number of SSs/GWs: [1 -10]

WiMAX link capacity: 4Mbps
4.3.5.2 Application Configuration

Server application is installed on the Base Station (BS).

Client application is installed on SSs.

Client application will send 900 messages, which have 20 bytes length, to
the Server application by using the Internet protocol UDP.

The Client application’s Data-Rate property is set to 900 x 20 bytes x 8
bits = 144,000bps = 144kbps
4.3.5.3 Simulation Planning

Repeatedly running the simulation scenarios with the different number of
SSs/GWs

Output: the network throughput, Tx delay
4.3.5.4 Results Analysis and Conclusion

Calculate the average network throughput, Tx delay
44

Conclusion: Do the BS receive all of the messages from the SSs/GWs in 1
second?
4.3.6 The WAN Simulation
4.3.6.1 Topology Configuration

Standard: Ethernet EEE 802.3

Connection mode: Point-To-Point

The BSs are connected to the Hub (or Data Center) in the Star topology

Number of BS: [1-20]

Ethernet link capacity: 10Mbps
4.3.6.2 Application Configuration

Server application is installed on the Hub (or DC)

Client application is installed on BSs.

Client application will send 9,000 messages, which have 20 bytes length,
to the Server application by using the Internet protocol UDP.

The Client application’s Data-Rate property is set to 9,000 x 20 bytes x 8
bits = 1,440,000bps = 1.44Mbps
4.3.6.3 Simulation Planning

Repeatedly running the simulation scenarios with the different number of
BSs

Output: the network throughput, Tx delay
4.3.6.4 Results Analysis and Conclusion

Calculate the average network throughput, Tx delay
45

Conclusion: Do the DC receive all of the messages from the BSs in 1
second?
4.4 Implementing the Network Simulation
4.4.1The WLAN Simulation
4.4.1.1 The NS-3 Script

Name: sm-ap-sim.cc

Description: This script implements the network model that simulates the
AMI meter data reporting process in a WLAN topology. The simulation
scenarios have one Wi-Fi Access Point (AP) and a number of the smart
meters (SM). The network devices are layout in a grid topology. The AMI
meter data reporting application will send the meter messages from the
SMs to the AP.

The source code of this script is in the Appendix session.

Syntax:
o Input:
nbSM - number of smart meter nodes to create [1]
duration - duration of the simulation in seconds [10]
verbose - turn on all WimaxNetDevice log components [false]
data-rate - packet data rate [0.160kbps]
statistic-start - the statistic is started at (second) [0]
o Output:
46
In every second:
Transmit (Tx) Packets, Receive (Rx) Packets, and
Maximum Tx Delay
In simulation period:
Average Transmit (Tx), Receive (Rx), and Transmit Delay
(TxDelay)
4.4.1.2 The Linux Shell Script

Name: sm-ap-sim.sh

Description: Batch running the WLAN simulation application. This shell
script generates many WLAN simulation scenarios. Then it simulates the
scenarios, and logs the simulation results in the text files.

Syntax:

Input: none

Output: List of the log file names that store the simulation results
4.4.2 The WNAN Simulation
4.4.2.1 NS-3 Script

Name: ap-gw-sim.cc

Description: This script implements the network model that simulates the
AMI meter data reporting process in a WNAN topology. The simulation
scenarios have one WiMAX/Wi-Fi gateway and a number of the mesh
routers. The network devices are layout in a grid topology. Some of the
47
mesh routers are configured as the APs. The AMI meter data reporting
application will send the meter messages from the APs to the gateway.
The source code of this script is in the Appendix session.

Syntax:
o Input:
x-size - number of columns of the grid [3]
y-size - Number of rows of the grid [3]
step - distance between two adjacent nodes (meter) [190]
access-points - number of Wi-Fi APs [1]
data-rate - packet data rate [20kbps]
statistic-start - the statistic is started at (second) [0]
o Output:
In every second:
Transmit (Tx) Packets, Receive (Rx) Packets, and
Maximum Tx Delay
In simulation period:
Average Transmit (Tx), Receive (Rx), and Transmit Delay
(TxDelay)
4.4.2.2 Linux Shell Script

Name: ap-gw-sim.sh

Description: Batch running the WNAN simulation application. This shell
script generates many WNAN simulation scenarios. Then it simulates the
scenarios, and logs the simulation results in the text files.
48

Syntax:
o Input: none
o Output:
List of the log file names that store the simulation results
4.4.3The WMAN Simulation
4.4.3.1 NS-3 Script

Name: gw-bs-sim.cc

Description: This script implements the network model that simulates the
AMI meter data reporting process in a WMAN topology. The simulation
scenarios have one WiMAX Base Station and a number of the Subscriber
Stations (or WiMAX/Wi-Fi Gateways). The network devices are layout in
a grid topology. The AMI meter data reporting application will send the
meter messages from the Subscriber Stations to the Base Station.
The source code of this script is in the Appendix session.

Syntax:
o Input:
nbSS - number of subscriber station to create [1]
scheduler - type of scheduler to use with the network devices [0]
duration - duration of the simulation (second) [10]
verbose - turn on all WimaxNetDevice log components [false]
data-rate - packet data rate [144kbps]
statistic-start - statistic started at (second) [0]
o Output:
49
In every second: Transmit (Tx) Packets, Receive (Rx) Packets, and
Maximum Tx Delay
In simulation period: Average Transmit (Tx), Receive (Rx), and
Transmit Delay (TxDelay)
4.4.3.2 Linux Shell Script

Name: gw-bs-sim.sh

Description: Batch running the WMAN simulation application. This shell
script generates many WMAN simulation scenarios. Then it simulates the
scenarios, and logs the simulation results in the text files.

Syntax:
o Input: none
o Output: List of the log file names that store the simulation results
4.4.4The WAN Simulation
4.4.4.1 NS-3 Script

Name: bs-dc-sim.cc

Description: This script implements the network model that simulates the
AMI meter data reporting process in a MAN topology. The simulation
scenarios have one Hub and a number of the WiMAX Base Stations. The
network devices are layout in a star topology. The AMI meter data
reporting application will send the meter messages from the Base Stations
to the Hub node (or the Data Center).
The source code of this script is in the Appendix session.
50

Syntax:
o Input:
nbBS - number of base station to create [1]
duration - duration of the simulation (second) [10]
verbose - turn on all WimaxNetDevice log components [false]
data-rate - packet data rate [1.44Mbps]
statistic-start - statistic started at (second) [0]
o Output:
In every second: Transmit (Tx) Packets, Receive (Rx) Packets, and
Maximum Tx Delay
In simulation period: Average Transmit (Tx), Receive (Rx), and
Transmit Delay (TxDelay)
4.4.4.2 Linux Shell Script

Name: bs-dc-sim.sh

Description: Batch running the WAN simulation application. This shell
script generates many WAN simulation scenarios. Then it simulates the
scenarios, and logs the simulation results in the text files.

Syntax:
o Input: none
o Output: List of the log file names that store the simulation results
Chapter 5
Simulation Results and
Analysis
5.1 The Simulation Experiments
We has conducted four types of the network simulation experiments based on the type of
network topologies, including WLAN, WNAN, WMAN, and WAN topologies. We have
run each simulation experiment at least five times to compare the simulation results.
After the results were validated, the results from the last simulation were used to prepare
the input for a new simulation cycle of the parent network in the hierarchy. For example,
the WLAN simulation shows that there are 100 UDP packets transmitted in one second
from the SMs to the AP. Then, in the parent network, WNAN, the APs will be configured
to send the same number of packets received from the WLAN simulation, or 100 packets
in this example.
5.1.1 The WLAN Simulation Experiment
52
The goal of the experiments is to evaluate the network throughput and the UDP packet
transmission delay with the different number of SMs in the WLAN topologies.
There are 11 experiments conducted to simulate the meter data traffic from the SMs to
the AP. The number of SMs is different between the scenarios, and between 1 and 100.
The location of the AP is fixed, but the location of the SMs was generated randomly in
each scenario. Based on the analysis of the Colorado Springs Utility network described
in Chapter 4, we observed there are average 50 SMs within 100x100 meter square area.
That is the reason we choose 50 as the number of SMs in a WLAN topology for the
simulation. The simulations results are shown and discussed in Section 5.3.1.
5.1.2 The WNAN Simulation Experiment
The goal of the experiment is to evaluate the network throughput and the UDP packet
transmission delay with the different number of MRs and APs in the WNAN topologies.
There are 9 experiments conducted to simulate the meter data traffic from the APs to the
GW. The number of APs is different between the experiments, and from 1 to 9. The
network devices, GW and MRs/APs were installed in a grid topology with the distance
between nodes is 200 meters. GW position is at the left-top node. MRs/APs are installed
at other nodes of grid. The number of hops in a mesh routing path will increase with the
increasing number of the MSs/APs. The NS-3 mesh simulation module which we used in
this study limits the number of APs to nine. The simulations results are shown and
discussed in Section 5.3.2.
5.1.3 The WMAN Simulation Experiment
53
The goal of the experiment is to evaluate the network throughput and the UDP packet
transmission delay with the different number of GWs in the WMAN topologies.
There are 10 experiments conducted to simulate the meter data traffic from the GWs to
the BS. The number of GWs is different between the experiments, and from 1 to 10. The
network devices, BS and GWs were installed in a grid topology with the distance
between nodes is 1,000 meters. BS position is at the left-top node. GWs are installed at
other nodes of grid. The NS3 WiMAX simulation module which we used in this study
limits the number of GWs to 20 but we have observed if the number of GWs increases
beyond ten, the packets will be lost.
Improve Meter Data Transmission Through Packet Aggregation
To improve the meter data transmission in the WMAN, we observed that the WiMAX
frame duration length is about 5 milliseconds that causes the maximum number of frames
processed in one second at a WiMAX/Wi-Fi gateway is about 200. It is not a good
utilization because there is only one meter data packet put in the sending UDP packet
during 5 ms of frame processing time. Instead, we can send more than one meter data
packets in one sending UDP packet by aggregating the received meter data packets at the
gateway into a single UDP packet and transmitting it with a WiMAX frame. For
example, in 5ms duration, the WiMAX connection with a transmission speed of 1Mbps
can deliver 5,000 bits or 625 bytes. If the length of the meter UDP packet is 20 bytes,
then the number of meter packets can be transmitted in one second is 31.
To evaluate the proposed improvement, we conduct simulation experiments , where 10
gateways connected to the base station in WiMAX point-to-multipoint mode. The
network performance is measured against the number of meter data packets put in a
54
WiMAX frame. The simulation parameter “number of meter data packets” increased until
the network is overloaded, or the number of received packets less than the number of sent
packets. The simulation results are shown and discussed in Section 5.3.3.
5.1.4 The WAN Simulation Experiment
The goal of the experiments is to evaluate the network throughput and the UDP packet
transmission delay with the different number of BSs in the WAN topologies.
There are 7 experiments conducted to simulate the meter data traffic from the BSs to the
DC. The number of BSs is different between the experiments, and between 1 and 7. The
network devices, DC and BSs, were installed in a star topology. The connection between
DC and BS is Point-to-Point that simulates that optical fiber connection. To evaluate the
affect of the cable length to the transmission delay of a UDP packet, we conduct 11
experiments that have the cable length changed from 1km to 100km. The simulations
results are shown and discussed in Section 5.3.4.
5.2 Simulation Data Collection
Because the network infrastructure simulation process was divided into four subnetworks such as WLAN, WNAN, WMAN, and WAN simulations. We can orderly
simulate and analyze each type of network topology. Therefore, we collected the
simulation results in each sub-network simulation. The simulation results were displayed
on the standard output device by the NS-3 C++ scripts. Four Linux shell scripts were
developed to run the simulation experiments many times for result validation. We have
modified the original standard output into the text file for offline further analysis.
55
5.3 The Simulation Results
The following tables show the simulation results. The specification of the simulation
design, and the NS-3 simulation implementation are included in Chapter 4.
5.3.1 The WLAN Simulation Results
5.3.1.1 Experiment 1: WLAN topology with 50 SMs
Topology configuration

Standard: Wi-Fi IEEE 802.11a

Connection mode: Infrastructure

The number of smart meters which the Wi-Fi AP serves: 50

Wi-Fi link capacity: 24Mbps

Smart meter location: random position within the coverage area of the AP
NS3
Simulation
Time
in
Sec
1
2
3
4
5
6
7
8
9
10
11
12
13
14
#
of
Packets
0
0
50
50
50
50
50
50
50
50
50
50
50
50
Tx # of Rx
Packets
0
0
50
50
50
50
50
50
50
50
50
50
50
50
Avg Tx Delay (µs)
0
0
36,011
9,800
10,025
10,211
9,418
9,581
9,164
10,685
9,587
10,023
9,993
9,835
# of Tx Packets / Total Processing
Meter
Delay (µs)
1
0
1
0
1
36,011
1
9,800
1
10,025
1
10,211
1
9,418
1
9,581
1
9,164
1
10,685
1
9,587
1
10,023
1
9,993
1
9,835
56
15
16
17
18
19
20
50
50
50
50
50
50
50
50
50
50
50
50
10,038
10,735
9,701
9,940
9,805
11,209
1
1
1
1
1
1
10,038
10,735
9,701
9,940
9,805
11,209
Table 5.13 – The WLAN simulation results with 50 SMs. The statistical data are in
one second period
The first two second periods of the simulation were in the initialization phase of the WiFi network infrastructure mode. In the initialization period, there were no data sent in the
first two seconds. Moreover, the third second period shows that the average delay to be
36,011 µseconds. This is due to the Wi-Fi nodes need to resolve the AP’s IP address
before they send the UDP meter data packets to it. Otherwise, the average delay is
converged to about 10,000 µseconds.
5.3.1.2 Summary of the WLAN Simulation Experiments
Avg. Tx
#
of # of Tx # of Rx Delay
Meters
Packets Packets (µs)
1
1
1
156
10
10
10
1,420
20
20
20
3,127
30
30
30
5,326
40
40
40
7,479
50
50
50
9,985
60
60
60
12,367
70
70
70
14,559
80
80
80
16,866
90
90
90
19,371
100
100
100
21,132
# of Tx Total
Packets / Processing
meter
Delay (µs)
1
156
1
1,420
1
3,127
1
5,326
1
7,479
1
9,985
1
12,367
1
14,559
1
16,866
1
19,371
1
21,132
Table 5.14 - The results of WLAN simulations with different number of SMs. The
statistical data in each row are the results of one simulation experiment.
5.3.2 The WNAN Simulation Results
57
5.3.2.1 Experiment 1: The WNAN topology with 9 APs
Topology Configuration

Standard: Wi-Fi IEEE 802.11s

Connection mode: Mesh

Wi-Fi link capacity: 54Mbps

The Mesh Routers (MR) /Access Points (AP) are installed in a grid topology

NS-3
Simulation
Time
in
Sec
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Distance between adjacent nodes (horizontal and vertical): 200 meters
# of Tx
Packets
441
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
# of Rx
Packets
374
405
400
412
450
265
374
303
456
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
Avg Tx
(µs)
5,037
8,031
730
3,718
685
4,959
234,084
189
5,507
224
213
227
203
198
206
208
213
198
264
219
199
191
190
251
Delay # of Tx Packets /
APs
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
Total Processing
Delay (µs)
251,839
401,555
36,518
185,894
34,239
247,975
11,704,216
9,442
275,363
11,207
10,636
11,366
10,171
9,899
10,295
10,391
10,665
9,886
13,223
10,960
9,959
9,538
9,490
12,550
58
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
293
290
266
239
246
214
152
158
130
130
277
245
240
236
236
5,472
146
148
146
146
2,828
192
192
200
191
268
176
180
179
176
243
206
190
189
192
205
2,937
227
231
233
231
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
14,637
14,475
13,308
11,950
12,281
10,677
7,591
7,901
6,478
6,524
13,851
12,229
12,010
11,809
11,786
273,620
7,289
7,396
7,287
7,308
141,386
9,621
9,608
9,980
9,539
13,390
8,824
9,001
8,933
8,789
12,125
10,297
9,493
9,440
9,610
10,266
146,842
11,348
11,553
11,640
11,533
59
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
450
449
450
450
450
450
450
450
450
450
5,518
246
237
233
236
269
170
175
164
171
1,340
159
161
161
165
250
214
221
215
223
2,877
254
244
247
254
2,881
156
151
149
148
173
1,206
128
130
130
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
275,912
12,305
11,835
11,631
11,784
13,467
8,485
8,731
8,190
8,526
66,994
7,960
8,066
8,049
8,243
12,522
10,692
11,050
10,742
11,151
143,872
12,717
12,186
12,354
12,718
144,026
7,808
7,559
7,440
7,378
8,667
60,298
6,397
6,499
6,479
Table 5.15 – The simulation results for WNAN 3x3 grid topology with nine APs. The
statistical data are in one second period
The initialization phase of the Wi-Fi Mesh network in the simulation occurred in the first
nine seconds. The mesh routing protocol needs to construct the routing path between the
60
APs and the gateway before UDP packets can be delivered to the receiver or mesh
gateway. This explains why there were the lost packets in the network initialization
period.
Other observation is the big jump on the delay of packet transmission at some simulation
seconds, i.e. 40, 45, 66. That is due to the changing in the routing path of the mesh
network. The new routing paths have more hops than the former ones. As a result, the
delay time has rapidly increased with the hop number on the routing path [HWC08].
In the first second period, the number of sent UDP packets was less than 450. That is due
to the schedule for starting of traffic applications. The start time of traffic applications on
the APs were extendedly shifted for the performance measurement. The traffic
applications that have the starting time shifted far away from the starting of the first
simulation second, could not send all 50 UDP packets in the first second as planning. As
a result, the total sent packets in the first second were less than 450.
The delay time was converged to 500 µseconds.
5.3.2.2 Summary of the WNAN Simulation Experiments
# of
APs
1
2
3
4
5
6
7
8
9
#
of
Packets
50
100
150
200
250
300
350
400
450
Tx #
of
Packets
50
100
150
200
250
300
350
400
450
Total
Rx Avg. Tx Delay # of Tx Packets / Processing
(µs)
APs
Delay (µs)
0
50
0
21
50
1,036
29
50
1,432
34
50
1,696
64
50
3,203
103
50
5,162
113
50
5,649
158
50
7,879
496
50
24,800
Table 5.16 – The results of WNAN simulation. The statistical data in each row are
the results of one simulation experiment.
61
5.3.3 The WMAN Simulation Results
5.3.3.1 Experiment 1: The WMAN topology with 10 GWs
Topology Configuration

Standard: WiMAX IEEE 802.16d

Connection mode: Point-To-Multipoint

WiMAX link capacity: 4Mbps

The Subscribers (SS)/Gateways (GW) are installed in the grid topology

Distance between adjacent nodes (horizontal and vertical): 1,000
meters
NS-3
Simulation
Time
in
Sec
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
# of Tx
Packets
0
0
0
0
0
0
1791
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
# of Rx
Packets
0
0
0
0
0
0
1782
1803
1803
1800
1802
1799
1797
1793
1802
1799
1804
1801
1802
Avg Tx
(µs)
0
0
0
0
0
0
5,315
5,266
5,312
5,372
5,348
5,297
5,305
5,315
5,315
5,295
5,290
5,368
5,277
Delay # of Tx Packets /
GW
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
Total Processing
Delay (µs)
0
0
0
0
0
0
956,704
947,964
956,076
966,898
962,597
953,517
954,901
956,689
956,619
953,111
952,218
966,236
949,810
62
20
21
22
23
24
25
26
27
28
29
30
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1800
1803
1798
1801
1798
1797
1796
1800
1802
1805
1800
1803
5,312
5,394
5,314
5,297
5,369
5,311
5,333
5,382
5,228
5,237
5,383
180
180
180
180
180
180
180
180
180
180
180
956,091
970,940
956,604
953,541
966,495
955,962
959,900
968,784
941,108
942,595
968,973
Table 5.17 – The simulation results for WMAN Point-to-multipoint topology with
ten GWs. The statistical data are in one second period.
There were no packets sent in the first six seconds as planning because of WiMAX
network initialization period. The number of sent packets in the seventh second was less
than 1,800 as planning because of the shifted starting time of traffic applications.
The IEEE 802.16d standard has the frame time of 5 milliseconds [IPGT10]. As a result,
the maximum number of WiMAX frames that can be sent in one second, is 180. The
average delay time of a UDP packet is close to the standard frame time.
5.3.3.2 Summary of the WMAN Simulation Experiments
# of
GWs
1
2
3
4
5
6
7
8
9
10
# of Tx
Packets
180
360
540
720
900
1080
1260
1440
1620
1800
# of Rx
Packets
180
360
540
720
900
1080
1260
1440
1620
1800
Avg. Tx Delay (µs)
5,147
5,158
5,171
5,166
5,161
5,159
5,166
5,158
5,229
5,321
Total
# of Tx Packets / Processing
GW
Delay (µs)
180
926,509
180
928,504
180
930,743
180
929,896
180
928,934
180
928,643
180
929,883
180
928,513
180
941,181
180
957,790
Table 5.18 – The WMAN simulation results. The statistical data in each row are the
results of a simulation experiment.
63
5.3.3.3 Summary of the Experiments for WMAN Improved
Design
#
of
Meter
Data
Packets
/
Tx
Packet
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Total
of Tx
Meter
# of Tx # of Rx Data
Packets
Packets Packets
1,800
1,800
1,800
1,800
1,800
3,600
1,800
1,800
5,400
1,800
1,800
7,200
1,800
1,800
9,000
1,800
1,800
10,800
1,800
1,800
12,600
1,800
1,800
14,400
1,800
1,800
16,200
1,800
1,800
18,000
1,800
1,800
19,800
1,800
1,800
21,600
1,800
1,800
23,400
1,800
1,800
25,200
1,800
1,800
27,000
1,800
1,786
28,800
1,800
1,786
30,600
Total
of Rx
Meter
Data
Packets
1,800
3,600
5,400
7,200
9,000
10,800
12,600
14,400
16,200
18,000
19,800
21,600
23,400
25,200
27,000
28,576
30,362
Tx Delay
(µs)
5,319
5,320
5,320
5,348
5,348
5,349
5,374
5,375
5,456
5,456
5,458
5,481
5,483
5,483
5,457
5,418
5,441
# of Tx
Packets /
Gateway
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
180
Total
Processing
Delay (µs)
957,337
957,608
957,608
962,608
962,714
962,879
967,363
967,469
982,120
982,120
982,419
986,613
986,912
986,985
982,330
975,151
979,464
Table 5.19 - The simulation results of a WMAN topologies experiment that has 10
gateways. The number of meter data packets put in a sending UDP packet was
changed until the UDP packet is fragmented.
In the above experiment, the number of UDP packets sent in one second is very close to
the maximum WiMAX frames can send in one second (about 200), that will cause the
network being overloaded when the UDP packets fragmented. For example, the number
of received packets is less than the number of sent packets when there are 16 or 17 meter
data packets in a UDP packet.
5.3.4 The WAN Simulation Results
5.3.4.1 Experiment 1: The WAN Star Topology With 3 BSs
Topology Configuration
64

Link capacity: 10Mbps

Medium Transmission Delay: 3.3 us/km for optical fiber, cable length is a
random number

BSs are connected to the Data Center in the Star topology

Distance between DC and BS: in range from 1 km to 100km

Connection mode: Point-to-point topology
NS-3
Simulation
Time
in
Sec
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
#
of
Packets
Tx
0
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
#
of
Packets
Rx
0
5,397
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
5,400
Avg Tx
(µs)
0
293
293
293
293
293
293
293
293
293
293
293
293
293
293
293
293
293
293
293
Delay # of Tx Packets /
BS
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
1,800
Total Processing
Delay (µs)
0
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
527,470
Table 5.20 – The simulation results of a WAN Point-to-point Star topology with
three base stations. The statistical data are in one second period.
65
The initialization period is in first two seconds. From the third second, the WAN network
has the average delay at 293 µseconds. The star topology has the point-to-point
connections between data center and base stations. This WAN network can provide a
very high bandwidth.
5.3.4.2 Summary of the Experiments for WAN Improved Design
#
of
Base
Stations
1
2
3
4
5
6
7
Tx
Packets
1,800
3,600
5,400
7,200
9,000
10,800
12,600
Rx
Packets
1,800
3,600
5,400
7,200
9,000
10,800
12,600
#
of
Meter
Data
Packets
/ Packet
15
15
15
15
15
15
15
Total Tx
Meter
Packets
27,000
54,000
81,000
108,000
135,000
162,000
189,000
Total Rx
Meter
Packets
27,000
54,000
81,000
108,000
135,000
162,000
189,000
Avg. Tx
Delay
(µs)
260
291
293
293
252
248
214
# of Tx
Packets /
BS
1,800
1,800
1,800
1,800
1,800
1,800
1,800
Total
Processing
Delay (µs)
468,070
524,500
527,470
527,470
453,814
447,280
385,756
Table 5.21 – The WAN simulation results. The statistical data in each row are the
results of one simulation experiment.
Cable
Length
(km)
1
10
20
30
40
50
60
70
80
90
100
#
of
Meter
# of Rx Data
# of Tx Packet Packets /
Packets s
Tx Packet
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
1,800
1,800
15
Total
Tx Total
Rx Avg.
Tx
Meter
Meter
Delay
Data
Data
Packets
Packets
(µs)
27,000
27,000
6
27,000
27,000
36
27,000
27,000
69
27,000
27,000
102
27,000
27,000
135
27,000
27,000
168
27,000
27,000
201
27,000
27,000
234
27,000
27,000
267
27,000
27,000
300
27,000
27,000
333
Total
Processing
Delay (µs)
10,690
64,150
123,550
182,950
242,350
301,750
361,150
420,550
479,950
539,350
598,750
Table 5.22 - The of simulation results of the WAN topologies experiments that have
one base station connected to the data center. The length of the optical fiber cable
was changed to evaluate the total processing delay at a BS.
66
5.4 Simulation Results Analysis
5.4.1 The WLAN Simulation Results Analysis
The Figure 5.1 show the simulation results in every one NS-3 second for the WLAN
topology that has 50 smart meters. Fifty UDP packets that sent from fifty smart meters,
received at the AP with the average transmission delay of 10 milliseconds.
We can see that, the delay time of a package and the total transmission delay converges
to 10 milliseconds.
WLAN (IEEE 802.11a) w/ 50 SMs Simulation
55
12,000
50
10,000
45
40
Packets
30
6,000
25
20
4,000
15
10
2,000
5
0
0
0
2
4
6
8
10
12
14
16
18
20
22
NS-3 Sim Seconds
Packets Tx
Packets Rx
Avg Tx Delay (us)
Total Processing Delay (us)
Figure 5.11 - The simulation results in every one second for the WLAN
Infrastructure mode topology with one AP and fifty SMs in the network.
Time (us)
8,000
35
67
120
25,000
100
20,000
Packets
80
15,000
60
10,000
40
Time (us)
WLAN (IEEE 802.11a) w/ variable SMs
Simulation
5,000
20
0
0
0
20
40
60
80
100
120
SMs
Packets Tx
Packets Rx
Total Processing Delay (us)
Figure 5.12 – The simulation results for the WLAN infrastructure mode topology.
The number of SMs is assigned from the one to one hundred in the experiments to
evaluate the changing of the total processing delay at the SMs.
Figure 5.2 shows the simulation results for the WLAN infrastructure mode topology. The
number of SMs is assigned from the one to one hundred in the experiments to evaluate
the changing of the total processing delay at the SMs. The number of the UDP packets
sent and received in every one second for the simulation duration versus the number of
the SMs in many different simulation scenarios. The total processing delay is also plotted
on the chart.
We can see that, the network has successfully transmitted the UDP packets in every one
second from the senders (or smart meters) to the receiver (or AP).
Moreover, we can see that the total processing delay increases linearly with the number
of smart meters. The total processing delay at one smart meter is below the one second
68
threshold when the number of smart meters in the WLAN topology is equal or less than
seventy.
5.4.2 The WNAN Simulation Results Analysis
WNAN (IEEE 802.11s Mesh) w/ 9 APs
Simulation
500
300,000
450
250,000
400
200,000
300
250
150,000
200
Time (us)
Packets
350
100,000
150
100
50,000
50
0
0
0
20
40
60
80
100
NS-3 Sim Seconds
Packets Tx
Packets Rx
Avg Tx Delay (us)
Total Processing Delay (us)
Figure 5.13 – The simulation results in every one second for the WNAN 3x3 mesh
topology with nine APs and one GW. AP sent 50 packets in every one second to the
GW.
Figure 5.14 shows simulation results in every one second for the WNAN 3x3 mesh
topology with nine APs and one GW. AP sent 50 packets in every one second to the GW.
There are total 450 UDP packets sent to the GW in every second. The number of sent
and received UDP packets is equal in every second. We also see that, the total
transmission delay converge to 500 millisecond.
69
WNAN (IEEE 802.11s Mesh) w/ variable APs
Simulation
500
30,000
25,000
400
300
15,000
200
Time (us)
Packets
20,000
10,000
100
5,000
0
0
0
1
2
3
4
5
6
7
8
9
APs
Packets Tx
Avg. Packets Rx
Total Processing Delay (us)
Figure 5.15 - The simulation results of the WNAN 3x3 grid topology with nine
MRs and one GW in the network.
Figure 5.4 shows the simulation results for the WNAN 3x3 grid topology. There are nine
MRs and one GW in the network. The number of APs is assigned from the one to nine in
the experiments to evaluate the changing of the total processing delay at the APs.
The number of the UDP packets sent and received in every one second for the simulation
duration versus the number of the APs in many different simulation scenarios. The total
processing delay is also plotted on the chart.
We can see that, the network has successfully transmitted the UDP packets in every one
second from the senders (or APs) to the receiver (or GW).
70
Moreover, we can see that the total processing delay increases rapidly with the number of
APs. The line of total processing time likes a parabolic arc. That is due to the increasing
hop count on the mesh routing path [AWW05, HWC08]. The total processing delay at
one AP for sending fifty 20-bytes packets is below the one second threshold.
5.4.3 The WMAN Simulation Results Analysis
1810
1,200,000
1805
1,000,000
1800
800,000
1795
600,000
1790
400,000
1785
200,000
1780
Time (us)
Packets
WMAN (IEEE 802.16d) w/ 10 GWs Simulation
0
0
5
10
15
20
25
30
35
NS-3 Sim Seconds
Packets Tx
Packets Rx
Avg Tx Delay (us)
Total Processing Delay (us)
Figure 5.16 - The simulation results of WMAN point-to-multipoint topology with
one BS and ten GWs in the network. GW sent 180 packets in every one second to
the BS.
Figure 5.5 shows the simulation results in every one second for the WMAN point-tomultipoint topology. There are one BS and ten GWs in the network. GW sent 180 packets
in every one second to the BS. There are 1,800 UDP packets sent to the BS in every
second. The number of received packets are not exactly equal the number of sent packets
in every second. The difference between them is between -5 and 5 packets. If a packet
71
does not arrive in the same second, it will arrive in the next second. The average delay of
a packet that is about 5.5 milliseconds, can validate this hypothesis.
We also see that, the total transmission delay converges to 930 milliseconds.
2,000
1,800
1,600
1,400
1,200
1,000
800
600
400
200
0
1,000,000
990,000
980,000
970,000
960,000
950,000
940,000
930,000
920,000
0
1
2
3
4
5
6
7
8
9
Time (us)
Packets
WMAN (IEEE 802.16d) w/ variable GWs
Simulation
10
GWs
Packets Tx
Avg. Packets Rx
Total Processing Delay (us)
Figure 5.17 - The simulation results for the WMAN topology. The number of GWs
is assigned from the one to ten in the experiments to evaluate the changing of total
processing delay at the GWs.
In the Figure 5.6 , the number of the UDP packets sent and received in every one second
for the simulation duration versus the number of GWs for different simulation
experiments. The total processing delay is also plotted on the chart.
We can see that, the network has successfully transmitted the UDP packets in every one
second from the senders (or GWs) to the receiver (or BS).
Moreover, we can see that the total processing delay is between 930 and 960
milliseconds. It does not increase with the number of the GWs. This is due to the
72
WiMAX network or 802.16d standard has a fixed frame time (5ms) that is independent to
the number of the subscribers.
Impact On The Network Performance By Aggregating Meter Data
Figure 5.7 shows the simulation results for the WMAN point-to-multipoint topology from
many experiments. There was one BS and ten GWs in the network. GW sent 180 packets
to the BS in every second. The number of meter data packets put in a transmitted UDP
packet was assigned from the one to seventeen packets to evaluate the changing of total
processing delay at the BSs in the experiments. Figure 5.7 shows the improvement of
network performance when
the number of meter data packets are aggregated in a
transmitted UDP packet or the length of loaded data in one WiMAX frame. The number
of meter data packet was increased until the network going to the overloaded state. As we
can see on the chart, when the number of meter packets is less than 16, the network
successfully transmitted all of the UDP packets. This is due to the sending UDP packet
that contains a designated number of meter data packets, is not fragmented in the
transmission. Moreover, the traffic application was programmed to send out in every
second the number of UDP packets that can be delivered completely by the network in
one second.
However, the network is overloaded when the number of embedded meter data packets is
equal or greater than 16. This is due to the UDP packets were fragmented in the
transmission. That caused number of received packets in one second less than the number
of sent packets. As a result, the average transmission delay of a packet was increased.
73
35,000
990,000
30,000
985,000
25,000
980,000
20,000
975,000
15,000
970,000
10,000
965,000
5,000
960,000
0
955,000
0
2
4
6
8
10
12
14
16
Time (us)
Packets
WMAN (IEEE 802.16d) w/ 10 GWs, various
sub-packet lengths simulation
18
Number of meter data packets
Packets Tx
Packets Rx
Total Sub Packet Tx
Total Sub Packet Rx
Total Processing Delay (us)
Figure 5.18 - Impact on the network performance by aggregating meter data.
5.4.4 The WAN Simulation Results Analysis
74
WAN (Star Topology) w/ 3 BSs Simulation
5,401
600,000
5,400
500,000
5,400
400,000
5,399
300,000
5,398
200,000
5,398
100,000
5,397
5,397
0
0
5
10
15
20
25
NS-3 Sim Seconds
Packets Tx
Packets Rx
Avg Tx Delay (us)
Total Tx Delay (us)
Figure 5.19 - The simulation results in every one second for the WAN star topology
with DC and three BSs in the network. BS sent 1,800 packets in every one second to
the DC.
Figure 5.20 shows the simulation results in every one second for the WAN star topology.
There are one DC and three BSs in the network. The connection between the DC and BS
is point-to-point. BS sent 1,800 packets in every one second to the DC. There are total of
5,400 packets sent to the DC in every second. The number of sent and received UDP
packets are equal in every second. We also see that, the total transmission delay
converges to 520 milliseconds.
Time (us)
Packets
5,399
75
WAN (Star Topology) w/ various BSs simulation
200,000
600,000
180,000
500,000
160,000
400,000
120,000
100,000
300,000
80,000
200,000
60,000
40,000
100,000
20,000
0
0
0
1
2
3
4
5
6
7
8
BSs
Packets Tx
Packets Rx
Total Meter Packet Tx
Total Meter Packet Rx
Total Processing Delay (us)
Figure 5.21 - The simulation results for the WAN star topology from many
experiments.
Figure 5.22 shows the simulation results for the WAN star topology from many
experiments. The number of BSs was assigned from the one to seven to evaluate the
changing of total processing delay at the BSs in the experiments. The number of the UDP
packets sent and received in every one second for the simulation duration versus the
number of BSs for different simulation experiments. The total processing delay is also
plotted on the chart.
We can see that, the network has successfully transmitted the UDP packets in every one
second from the senders (or BSs) to the receiver (or the data center).
Time (us)
Packets
140,000
76
Moreover, we can see that the total processing delay is independent from the number of
the BSs. This is due to the BSs were connected to the data center in the point-to-point
connections. The network can transmit over 180,000 meter data packets that sent from
seven BSs, and the total processing time at each BS is less than 600 milliseconds.
However, the average delay is affected by the distribution of the BSs around the data
center.
WAN (Star Topology) w/ various cable lengths
simulation
30,000
700,000
25,000
600,000
Packets
400,000
15,000
300,000
10,000
200,000
5,000
100,000
0
0
0
20
40
60
80
100
120
Cable Length
Packets Tx
Packets Rx
Total Meter Packet Tx
Total Meter Packet Rx
Total Processing Delay (us)
Figure 5.23 - The simulation results for the WAN star topology with one DC and
one BS. The length of the optical fiber cable was assigned from the one to 100
kilometers.
Figure 5.24 shows the simulation results for the WAN star topology from many
experiments. There was one DC and one BS in the network. BS sent 1,800 packets to the
Time (us)
500,000
20,000
77
DC in every second. The length of the optical fiber cable that connects the DC and BS
was assigned from the one to 100 kilometers to evaluate the changing of total processing
delay at the BS in the experiments. We can see that the total processing time was linearly
increased with the length of the optical fiber cables that connect the BSs and the data
center.
Chapter 6
Lessons Learned
6.1 The Development of SG-WIP Planning Tool
SG-WIP web application is a Google Maps mashup. The developing process is followed
the steps Requirement Spec, Design Spec, Coding and Testing, Deployment.
One of the most challenges of the Web application development process is the debugging
task. The debugging web pages have to be opened in a Web browser that supports the
debugging tool, such as the Developer Tools in the Microsoft Internet Explorer.
However, the debugging tool of the Web browser can only debug the scripts that run on
the client machine, i.e. JavaScript. The Google Maps mashup SG-WIP uses PHP
language at the Web server and JavaScript at the client browser. When a web page is
running at the client browser, the server scripts of that web page have been executed on
the web server before it is sent to the browser. So the debugging tool at the web browser
can not be used for the server scripts. Because we can not interactively debug the server
scripts from the browser tools, we need to log the running information of the server
scripts for the program tracing. We also converted the PHP server pages into the
79
executable programs and ran them on the server to test and debug the server scripts much
easier.
Another difficulty is the collecting of the housing unit geographical locations for the
residential density analysis, and the street light poles geographical locations for the
planning of the wireless antennae. We have the helps of CSU to get needed infrastructure
information for planning the AMI communication network in the simulation process.
Although the infrastructure information collection was a sophisticate process because
there were constrains on the data release policy from CSU, the topology planning tool has
the real world geographical location of the housing units and the streetlight poles to
display and process. Thus, the simulation network model is designed closer to the real
world.
6.2 The Development of Smart Grid Simulation
Model
This thesis employed the modern wireless network technologies such as WiMAX and
Wi-Fi. Network simulation tools help the network researchers evaluating their solutions,
designing new network devices or communication protocols. We have used two network
simulation software NS-3 and NCTUns [NS3, NCTU01] in researching and developing
the WiMAX/Wi-Fi communication infrastructure for AMI.
NCTUns has a great GUI that allows visually building the network topologies,
simulating, and animating the network simulation. We used the NCTUns to quickly
construct the simulation scenarios in the thesis proposal development phase.
NS-3 was used to construct the complicated parameterized wireless network model in this
thesis. There are hundreds thousand of the wireless nodes (smart meters) simulated using
the NS-3 simulator. One of the greatest features of NS-3 is providing the programmable
80
network modules such as WiMAX, Wi-Fi, CSMA under the Object Oriented
Programming (OOP) style in C++ language.
However, NS-3 is still in developing phase. There are not much help documents
published. I have found a bug in the WiMAX module that limits the number of the
Subscriber Stations to under twenty nodes in any simulation scenario. More information
about
this
bug
is
following
the
URL:
http://www.nsnam.org/bugzilla/show_bug.cgi?id=1025
NS-3 is free and open source project. It has been quickly accepted and supported by the
network research community.
6.3 The Simulation Process in NS-3
6.3.1 The Initialization Phase of Wireless Networks
NS-3 wireless network models such as WiMAX, and Wi-Fi perform the Initialization
phase at the starting of simulation duration. The time length of the initialization phase is
dependent on the network type, for example, the WiMAX topology with one BS and ten
SSs needs six seconds.
The simulation experiment designs should take this fact into account when making
decision about the start time of the network traffic applications.
6.3.2 The Bugs in NS-3 Module Code
NS-3 is still in the development phase. Whether you like it or not, the bugs existing in the
source code, and they will affect to your simulation experiments.
As I mentioned about the bug in the WiMAX module, I has sent the bug report to the
authors of the WiMAX module and immediately received their responses.
81
However, the open source software like NS-3 has been rapidly developing and receiving
the supports from the network research and development community. When there is a
bug reported, there are hundreds of the researchers around the world who can help
solving it immediately. As a result, the open source software NS-3 simulator will be more
and more reliable and helpful.
Chapter 7
Simulation Limitations and
Future Work
The proposed network model is implemented as a parameterized simulator in NS-3.
There are many featured variables that were used as the input of the simulator, such as
number of network device nodes, distance between nodes, data rates, etc. However, there
are some limitations that were not modeled in the simulator, for example, the height of
the nodes, geographical locations.
7.1 Display Simulation Results on the SG-WIP
Planning Tool
In this thesis, the designed simulation experiences are based on the analysis of the
household density in the service areas. Because the real world network topologies are
complicated, in the simulation, we used only one featured simulation scenario designed
for a specific kind of the network topology i.e. WAN, WMAN, WNAN, and WLAN.
83
We can use our SG-WIP planning tool to edit the network topologies, simulation the trial
topologies, and display the simulation results on the GUI display. Then, we can compare
the different planned topologies to select the better one.
The SG-WIP planning tool can be extended to export the network topologies and
planning results including the placement of antennae as a file.
The NS-3 network model can be modified to accept an arbitrary network topology as an
input. Then, the real world topologies can be simulated with the NS-3 network model and
produce simulation results for further analysis.
7.2 Alternative Method
Application Simulation
for
Network
Traffic
The alternative method for application traffic simulation is the sending packet at a
random time. The packets are randomly sent from the smart meters to the data center
through the communication infrastructure. When the packets are transmitted in the
communication infrastructure, they will be logged in the trace files on the intermediate
network nodes along the routing path.
One of the advantages of this traffic simulation method is the randomly sending time of
the packets, which makes the simulation closer to the real world. However, in the
complicated simulation scenarios, such as in this thesis, the alternative method needs
more computing resources such as CPU time for random number generation, memory
space, storage space for trace files.
7.3 Improve the Antenna Placement Algorithm
84
The antenna placement algorithm in the SG-WIP application should take into account the
network availability property when it searches for the antenna position. For network
availability improvement, the antenna will cover not only the corresponding network
area, but also its neighborhood networks. If the antenna of a network is down, the
working antenna in one of the neighborhood network areas will become the
corresponding one.
7.4 Store the Real-time Meter Data in the Database
Management System (DBMS)
DBMS is needed to store the real-time meter data for efficient data access and
management. The real-time metering collection process may pass a large amount of
meter data to the DBMS in less than one second. It opens an opportunity for a research of
the DBMS for the real-time meter data. The possible solution can employ the real-time
DBMS technologies, and the distributed computing technologies.
7.5 Evaluate the Performance of the Network Model
with the AMI real-time Demand Response
Applications
Demand response is one of the important goals of the AMI deployment. In contrast to the
metering collection, the demand response supporting applications will request the meter
data from the data center for the consumer’s demand analysis. Then the demand response
applications can help the consumers optimize their energy usage. Although this subject is
out of the scope of this thesis, it can contribute to the AMI researcher community another
interesting performance evaluation of the WMN WiMAX/Wi-Fi infrastructure.
Chapter 8
Conclusion
AMI is being implemented by many utilities around the world. AMI contributes the
benefits not only to the utilities but also to the consumers. AMI real-time meter data
collection can give the utilities and consumers the ability to access the real-time meter
data. Consumers are benefits from the sharing real-time meter data because they can
monitor and actively adjust their demand of electric, gas, and water to save money.
Utilities are benefits from the real-time meter data because they can use the real-time
meter data to improve the quality of load profile charts, and load prediction. So the
utilities can save the fuel usage of power plants and reduce the price of electricity.
Many utilities have implemented the AMI wireless infrastructures for collecting meter
data automatically. However, most of the deployed wireless infrastructure did not support
or have not supported yet the real-time meter data collection. The intervals for meter data
collection are typically higher than one minute. The current meter data collecting period
is often in the range between fifteen and forty five minutes.
86
One of the main contributions of this thesis is the development a methodology for
measuring the performance of the hybrid WiMAX/Wi-Fi communication infrastructure
for the real-time metering data collection. The second contribution is the development of
a software tool SG-WIP for planning and designing the AMI wireless infrastructure using
the real utility light poles and meters GIS data from the city of Colorado Springs,
Colorado. The third contribution is the development of a simulator SG-SIM to evaluate
the performance
of the hybrid WiMAX/Wi-Fi AMI network. We proposed a
parameterized WiMAX/Wi-Fi network model and implemented it in the NS-3 platform.
Experiments were conducted using the network simulation process, including the WLAN
(Wi-Fi) simulation, the WNAN (Wi-Fi Mesh) simulation, the WMAN (WiMAX)
simulation, and the WAN (optical fiber point-to-point connection) simulation. The
simulation results show that the proposed WiMAX/Wi-Fi WMN infrastructure can
transport the meter data from 160,000 smart meters in the CSU service areas to the data
center in one second.
From the simulation result analysis, we can conclude that the high scalability property of
WiMAX/Wi-Fi WMN helps flexibly extend the coverage area of the AMI wireless
infrastructure without degrading the network performance.
This thesis provides the utilities an AMI wireless communication infrastructure solution
that employs the WiMAX/Wi-Fi WMN architecture for real-time metering data
collection. The proposed WiMAX/Wi-Fi infrastructure allows the utilities deploying an
AMI wireless communication infrastructure not only at low cost of installation and
maintenance but also with high performance, scalability, and security.
Bibliography
[DoE01] U.S. Department of Energy, “Smart Grid”,
<http://www.oe.energy.gov/smartgrid.htm>
[DoE02] U.S Department of Energy, “Smart Grid: An Introduction”,
<http://www.oe.energy.gov/SmartGridIntroduction.htm>
[Wiki01] “Smart Grid”, <http://en.wikipedia.org/wiki/Smart_grid>
[NIST10] National Institute of Standards and Technology, “NIST Framework and
Roadmap for Smart Grid Interoperability Standards, Release 1.0”, Jan. 2010.
[NETL08] National Energy Technology Laboratory, white paper “Advanced Metering
infrastructure”, February 2008.
[Chow09] Edward Chow, Lecture “Secure Smart Grids”, Department of Computer
Science, University of Colorado at Colorado Springs, 2009.
[IEEE11] IEEE Standard 802 Part 11: Wireless LAN Medium Access Control (MAC)
and Physical Layer (PHY) Specifications, 2007.
[IEEE15] IEEE Standard 802 Part 15.1: Wireless Medium Access Control (MAC) and
Physical Layer (PHY) Specifications for Personal Area Networks (WPANs), 2005.
[IEEE16] IEEE Standard 802 Part 16: Air Interface for Broadband Wireless Access
Systems, 2009.
[IEEE11s] IEEE, “Draft amendment: ESS mesh networking”, IEEE P802.11s Draft 1.00,
November 2006.
88
[Moh01] Prasant Mohapatra, Lecture “Wireless Mesh Networks”, Department of
Computer Science University of California, Davis.
[AWW05] I. F. Akyildiz, X. Wang, and W. Wang, "Wireless Mesh Networks: A
Survey," Computer Networks Journal (Elsevier), vol. 47, no. 4, pp. 445-487, Mar. 2005.
[Kri01] Srini Krishnamurthy, “Smart AMI Network Solutions Enable the Smart Grid”,
ElectricEnergyOnline.com,
<http://www.electricenergyonline.com/?page=show_article&mag=55&article=395>
[Met01] MetroFi, <http://en.wikipedia.org/wiki/MetroFi>
[Sky01] SkyPilot, <http://skypilot.trilliantinc.com>
[Eka01] EkaNet, <http://www.ekasystems.com/ekanet.htm>
[AWW05] I. F. Akyilidz, X. Wang, and W. Wang, “Wireless Mesh Networks: A Survey”
in Elsevier Computer Networks Journal, vol. 47, March 2005, pp. 445–487.
[JS03] J. Jangeun and M. L. Sichitiu, “The Nominal Capacity of Wireless Mesh
Networks,” in IEEE Wireless Communications Magazine, October 2003, vol. 10 no. 5,
pp. 8–14.
[RC05] A. Raniwala and T. cker Chiueh, “Architecture and Algorithms for an IEEE
802.11-based Multi-channel Wireless Mesh Network,” in Proceedings of INFOCOM
2005, March 2005, vol. 3, pp. 2223–2234.
[ANMK08] Akhtar, Nadeem and Moessner, Klaus, “Capacity of Grid-Oriented Wireless
Mesh Networks”, 3rd International Conference on Communication Systems Software and
Middleware and Workshops, Volumes 1 and 2 . pp. 631-636.
89
[HWC05] Jane-Hwa Huang, Li-Chun Wang, Chung-Ju Chang, “Coverage and capacity
of a wireless mesh network”, Wireless Networks, Communications and Mobile
Computing, 2005 International Conference on, Vol. 1 (2005), pp. 458-463.
[CK08] Joseph D. Camp and Edward W. Knightly, “The IEEE 802.11s Extended Service
Set Mesh Networking Standard”, IEEE Communications Magazine, Vol. 46, No. 8.
(August 2008), pp. 120-126.
[KSCV06] P. Kyasanur, J. So, C. Chereddi, and N. H. Vaidya ,”Multi-Channel Mesh
Networks: Challenges and Protocols”, in IEEE Wireless Communications, April 2006.
[IPGT10] Mohamed Amine Ismail, Giuseppe Piro, Luigi Alfredo Grieco, Thierry
Turletti, “An Improved IEEE 802.16 WiMAX Module for the ns-3 Simulator”,
Proceedings of SIMUTools Conference, 2010 , March, 2010.
[INTL04] Intel Corporation, white paper “Understanding Wi-Fi and WiMAX as MetroAccess Solutions”, 2004.
[LLT03] B. Liu, Z. Liu, and D. Towsley, "On the capacity of hybrid wireless networks",
in Proceedings of IEEE INFOCOM, Mar. 2003, vol. 2, pp. 1543-1552.
[ZR06] S. Zhao and D. Raychaudhuri, "On the Scalability of Hierarchical Hybrid
Wireless Networks, Proceedings of the Conference on Information Sciences and Systems
(CISS 2006), March 2006, pp. 711-716.
[ZSR04] S. Zhao, I. Seskar and D. Raychaudhuri, "Performance and Scalability of SelfOrganizing Hierarchical Ad-Hoc Wireless Networks," Proceedings of the IEEE Wireless
Communications and Networking Conference (WCNC'04), Atlanta, GA. March 2004,
pp. 132-137.
90
[HWC08] J. H. Huang, L. C. Wang, C. J. Chang, “Wireless Mesh Network: Architecture
and Protocols”, chapter title “Architectures and Deployment Strategies for Wireless Mesh
Networks”, Springer 2008.
[OSI] “OSI Model”, <http://en.wikipedia.org/wiki/OSI_model>
[Wimax] WiMAX community, <http://www.wimax.com>
[NS3] The Network Simulator Ns-3, <http://www.isi.edu/nsnam/ns/>
[NCTU01]
NCTUns
6.0
Network
Simulator
and
Emulator,
<http://nsl.csie.nctu.edu.tw/nctuns.html>
[NCTU02] “The Protocol Developer Manual for the NCTUns 6.0”, Network and System
Laboratory, Department of Computer Science, National Chiao Tung University, Taiwan
2010.
[HSWL07] S.M. Huang, Y.C. Sung, S.Y. Wang, and Y.B. Lin, “NCTUns Simulation
Tool for WiMAX Modeling,” Third Annual International Wireless Internet Conference,
October 22 – 24, 2007, Austin, Texas, USA. (EI and ISI indexed, sponsored by ICST,
ACM, and EURASIP)
[SH06] N.B. Salem and J.P. Hubaux, "Securing Wireless Mesh Networks," Wireless
Comm., vol. 13, no. 2, 2006, pp. 50–55.
[PSC06] Michael Purvis, Jeffrey Sambells, and Cameron Turner, “Beginning Google
Maps Applications with PHP and Ajax”, Apress, 2006.
[Goog01] Google Maps JavaScript V3,
<http://code.google.com/apis/maps/documentation/javascript/>
Appendix A
SG-WIP User Manual
A.1 Installation
A.1.1 System Requirements

Apache Web Server

PHP module

MySQL Server
A.1.2 The Database
The database is named “sgwip”. There is a user account name “sgwip” created for
accessing the database from the Web application “SG-WIP”. The “sgwip” account need
proper access permission to allow the Web application “SG-WIP” access the tables in the
database.
A.1.3 The SG-WIP Application
The
URL
of
the
SG-WIP
for
downloading
is
at
http://cs.uccs.edu/~gsc/pub/master/phuynh/src/sgwip.zip
The Web application SG-WIP is installed on the Apache Web server.
For example the application was install on the local web server as the following Linux
Fedora file system: /usr/share/sgwip
92
Figure A.7 - The SG-WIP application was installed on the Apache web server
93
A.2 The GUI Operations
A.2.1 Opening the Home Page
In the Web browser, open the following URL: http://scad.eas.uccs.edu/sgwip/wan.html
Figure A.8 - The home page of SG-WIP
94
A.2.2 Generating the WAN Topology

Click the menu item “New”

Wait for the GUI being generated and displayed on the Google maps (this process
may take many seconds to finish the job)
Figure A.9 - The WAN topology has twelve WiMAX base stations

Click the menu item “WAN”
95
Note: To return back to the WAN topology in the current path of the topology
exploration
A.2.3 Generating the MAN Topology

Click on the corresponding rectangle from the WAN topology GUI. A rectangle is a
Google Maps overlay object that represents for a network topology.

Wait for the network topology GUI being generated and displayed on the Google
Maps (this process may take many seconds to finish the job)

Click the menu item “MAN”
Figure A.10 - The WMAN topology has WiMAX BS, and WiMAX/Wi-Fi Gateways.
Note: To return back to the MAN topology in the current path of the topology exploration
96
A.2.4 Generate the NAN Topology

Click on the corresponding rectangle from the MAN topology GUI. A rectangle is a
Google Maps overlay object that represents for a network topology.

Wait for the network topology GUI being generated and displayed on the Google
maps (this process may take many seconds to finish the job)
Figure A.11 The WNAN topology has WiMAX/Wi-Fi gateway, and Mesh
routers/Access Points.

Click the menu item “NAN”
Note: To return back to the NAN topology in the current path of the topology
exploration
97
A.2.5 Generating the LAN Topology

Click on the corresponding rectangle from the NAN topology GUI. A rectangle is a
Google Maps overlay object that represents for a network topology.

Wait for the network topology GUI being generated and displayed on the Google
maps (this process may take many seconds to finish the job)
Figure A.12 - The WLAN topology includes Wi-Fi Access Point, and Smart meters.

Click the menu item “LAN”
98
Note: To return back to the LAN topology in the current path of the topology
exploration
A.2.6 Changing the Antennae Location of Network Devices
The antennae of the wireless network devices such as WiMAX base station,
WiMAX/Wi-Fi gateway, or AP/Mesh Router can be re-located to a better place within
the topology.

Click the menu item “On-Change-Antenna-Loc”

Click the changed Antenna

Navigate to the destination network topology where there is a hanging object such as
street light poles, or the housing/building unit.

Click the target hanging object to select a new location for the antenna.
Appendix B
SG-SIM Simulator
Running Examples
B.1 The LAN Simulation Examples
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/sm-ap-sim --nbSM=1 --statisticstart=4"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.273s)
1
0
0
0ns
2
0
0
0ns
3
1
1
6082144ns
4
1
1
1540048ns
5
1
1
1540048ns
6
1
1
1540048ns
7
1
1
1540048ns
8
1
1
1540048ns
9
1
1
1540048ns
10
1
1
1540048ns
Avg. Tx packets/second:1
Avg. Rx packets/second:1
Avg. Tx delay (milliseconds):1
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/sm-ap-sim --nbSM=10 --statisticstart=4"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.272s)
1
0
0
0ns
2
0
0
0ns
3
10
10
66400682ns
4
10
10
19814024ns
100
5
10
10
19277024ns
6
10
10
19915025ns
7
10
10
20013025ns
8
10
10
19392024ns
9
10
10
20553025ns
10
10
10
20255025ns
Avg. Tx packets/second:10
Avg. Rx packets/second:10
Avg. Tx delay (milliseconds):19
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/sm-ap-sim --nbSM=100 --statisticstart=4"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
[ 519/1090] cxx: examples/sg-sim/sm-ap-sim.cc -> build/debug/examples/sg-sim/sm-apsim_3.o
[1090/1090] cxx_link: build/debug/examples/sg-sim/sm-ap-sim_3.o
build/debug/examples/sg-sim/sg-onoff-application_3.o -> build/debug/examples/sgsim/sm-ap-sim
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (4.568s)
1
0
0
0ns
2
0
0
0ns
3
100
100
688476554ns
4
100
100
197641136ns
5
100
100
195392137ns
6
100
100
199219136ns
7
100
100
197660136ns
8
100
100
197337137ns
9
100
100
201163137ns
10
100
100
201171136ns
Avg. Tx packets/second:100
Avg. Rx packets/second:100
Avg. Tx delay (milliseconds):198
101
B.2 The NAN Simulation Examples
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/ap-gw-sim --data-rate=16kbps -time=30 --access-points=1 --x-size=1 --y-size=1 --interfaces=4 --step=300 --statisticstart=12"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.247s)
1
99
99
0ns
2
100
100
0ns
3
100
100
0ns
4
100
100
0ns
5
100
100
0ns
6
100
100
0ns
7
100
100
0ns
8
100
100
0ns
9
100
100
0ns
10
100
100
0ns
11
100
100
0ns
12
100
100
0ns
13
100
100
0ns
14
100
100
0ns
15
100
100
0ns
16
100
100
0ns
17
100
100
0ns
18
100
100
0ns
19
100
100
0ns
20
100
100
0ns
21
100
100
0ns
22
100
100
0ns
23
100
100
0ns
24
100
100
0ns
25
100
100
0ns
26
100
100
0ns
27
100
100
0ns
28
100
100
0ns
29
100
100
0ns
30
100
100
0ns
Avg. Tx packets/second:100
Avg. Rx packets/second:100
Avg. Tx delay (milliseconds):0
102
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/ap-gw-sim --data-rate=16kbps -time=30 --access-points=4 --x-size=2 --y-size=2 --interfaces=4 --step=300 --statisticstart=12"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.252s)
1
396
380
103982400ns
2
400
400
874410ns
3
400
400
856410ns
4
400
400
866410ns
5
400
400
902410ns
6
400
400
211025206ns
7
400
400
41414ns
8
400
400
41414ns
9
400
400
41414ns
10
400
400
41414ns
11
400
400
894996ns
12
400
400
41414ns
13
400
400
41414ns
14
400
400
41414ns
15
400
400
41414ns
16
400
400
1273064ns
17
400
400
41414ns
18
400
400
41414ns
19
400
400
41414ns
20
400
400
41414ns
21
400
400
1223822ns
22
400
400
41414ns
23
400
400
41414ns
24
400
400
41414ns
25
400
400
41414ns
26
400
400
1135480ns
27
400
400
41414ns
28
400
400
41414ns
29
400
400
41414ns
30
400
400
41414ns
Avg. Tx packets/second:400
Avg. Rx packets/second:400
Avg. Tx delay (milliseconds):0
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/ap-gw-sim --data-rate=16kbps -time=30 --access-points=6 --x-size=2 --y-size=3 --interfaces=4 --step=300 --statisticstart=12"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
103
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.272s)
1
594
570
106987685ns
2
600
600
1632058ns
3
600
600
1441058ns
4
600
600
1642057ns
5
600
600
1837642ns
6
600
372
205857651ns
7
600
828
1639371823ns
8
600
600
1255824ns
9
600
600
1337824ns
10
600
600
1255824ns
11
600
600
3000281ns
12
600
600
1382824ns
13
600
600
2284941ns
14
600
600
321064ns
15
600
600
321064ns
16
600
600
2713123ns
17
600
600
1274290ns
18
600
600
993955ns
19
600
600
319412ns
20
600
600
319412ns
21
600
600
989117ns
22
600
600
924711ns
23
600
600
2113878ns
24
600
600
319412ns
25
600
600
319412ns
26
600
600
2505356ns
27
600
600
808770ns
28
600
600
2939821ns
29
600
600
318997ns
30
600
600
318997ns
Avg. Tx packets/second:600
Avg. Rx packets/second:600
Avg. Tx delay (milliseconds):1
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/ap-gw-sim --data-rate=16kbps -time=30 --access-points=9 --x-size=3 --y-size=3 --interfaces=4 --step=300 --statisticstart=12"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.272s)
1
891
738
119623509ns
2
900
901
1001745189ns
104
3
900
900
2325664ns
4
900
900
2484352ns
5
900
853
4331459ns
6
900
420
207812067ns
7
900
652
1639814429ns
8
900
588
206033357ns
9
900
918
210843707ns
10
900
900
1398469ns
11
900
900
2194660ns
12
900
900
3773257ns
13
900
900
3320283ns
14
900
900
2380716ns
15
900
900
832062ns
16
900
900
4425319ns
17
900
900
2087537ns
18
900
900
3081115ns
19
900
900
2029719ns
20
900
900
1498443ns
21
900
900
3893531ns
22
900
900
2187714ns
23
900
900
2989236ns
24
900
900
955469ns
25
900
900
1240705ns
26
900
900
2895261ns
27
900
900
2384426ns
28
900
900
4603544ns
29
900
900
821468ns
30
900
900
853298ns
Avg. Tx packets/second:900
Avg. Rx packets/second:900
Avg. Tx delay (milliseconds):2
105
B.3 The MAN Simulation Examples
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/gw-bs-sim --nbSS=1 --duration=30
--statistic-start=10"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.291s)
1
0
0
0ns
2
0
0
0ns
3
0
0
0ns
4
0
0
0ns
5
0
0
0ns
6
0
0
0ns
7
900
892
12140304ns
8
900
901
11695508ns
9
900
901
11917504ns
10
900
902
12139374ns
11
900
900
12139064ns
12
900
901
12138628ns
13
900
901
11916180ns
14
900
891
11915853ns
15
900
910
12137866ns
16
900
892
12137472ns
17
900
902
11915024ns
18
900
900
12136936ns
19
900
901
11914387ns
20
900
901
11913951ns
21
900
902
12136048ns
22
900
900
12135696ns
23
900
900
12135260ns
24
900
901
12135034ns
25
900
892
11912460ns
26
900
901
11912276ns
27
900
901
12134104ns
28
900
902
12133794ns
29
900
900
11689065ns
30
900
901
11911061ns
Avg. Tx packets/second:900
Avg. Rx packets/second:900
Avg. Tx delay (milliseconds):12
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/gw-bs-sim --nbSS=5 --duration=30
--statistic-start=10"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
106
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.261s)
1
0
0
0ns
2
0
0
0ns
3
0
0
0ns
4
0
0
0ns
5
0
0
0ns
6
0
0
0ns
7
4496
4463
12279339ns
8
4500
4505
12278987ns
9
4500
4505
12278635ns
10
4500
4507
12193302ns
11
4500
4502
12278141ns
12
4500
4505
12277831ns
13
4500
4505
12192456ns
14
4500
4468
12277211ns
15
4500
4495
12276859ns
16
4500
4503
12191400ns
17
4500
4509
12191090ns
18
4500
4501
12275845ns
19
4500
4504
12275661ns
20
4500
4505
12190202ns
21
4500
4510
12274873ns
22
4500
4500
12052408ns
23
4500
4503
12137719ns
24
4500
4497
12188962ns
25
4500
4466
12273801ns
26
4500
4505
12051143ns
27
4500
4505
12188032ns
28
4500
4509
11913989ns
29
4500
4500
12272561ns
30
4500
4505
12272293ns
Avg. Tx packets/second:4500
Avg. Rx packets/second:4500
Avg. Tx delay (milliseconds):12
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/gw-bs-sim --nbSS=10 -duration=30 --statistic-start=10"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.268s)
1
0
0
0ns
2
0
0
0ns
3
0
0
0ns
107
4
0
0
0ns
5
0
0
0ns
6
0
0
0ns
7
8991
8935
20416344ns
8
9000
9007
20415908ns
9
9000
9002
20415724ns
10
9000
9024
20415372ns
11
9000
9004
20415062ns
12
9000
8994
20414626ns
13
9000
8991
20414484ns
14
9000
8951
20414132ns
15
9000
9010
20413864ns
16
9000
9007
20413470ns
17
9000
9013
20413244ns
18
9000
9007
20412892ns
19
9000
9007
20412582ns
20
9000
9010
20412314ns
21
9000
9009
20412004ns
22
9000
8999
20411694ns
23
9000
8982
20411342ns
24
9000
8998
20411116ns
25
9000
8965
20410680ns
26
9000
9009
20410496ns
27
9000
9011
20410186ns
28
9000
9013
20409792ns
29
9000
9004
20409566ns
30
9000
9010
20409214ns
Avg. Tx packets/second:9000
Avg. Rx packets/second:9000
Avg. Tx delay (milliseconds):20
108
B.4 The WAN Simulation Examples
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/bs-dc-sim --nbBS=3 --duration=10
--statistic-start=3 --data-rate=1.44Mbps"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.261s)
1
0
0
0ns
2
27000
26943
2079999ns
3
27000
27000
2079999ns
4
27000
27000
2079999ns
5
27000
27000
2079999ns
6
27000
27000
2079999ns
7
27000
27000
2079999ns
8
27000
27000
2079999ns
9
27000
27000
2079999ns
10
27000
27000
2079999ns
Avg. Tx packets/second:27000
Avg. Rx packets/second:27000
Avg. Tx delay (milliseconds):2
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/bs-dc-sim --nbBS=5 --duration=10
--statistic-start=3 --data-rate=1.44Mbps"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.266s)
1
0
0
0ns
2
45000
44905
2079999ns
3
45000
45000
2079999ns
4
45000
45000
2079999ns
5
45000
45000
2079999ns
6
45000
45000
2079999ns
7
45000
45000
2079999ns
8
45000
45000
2079999ns
9
45000
45000
2079999ns
10
45000
45000
2079999ns
Avg. Tx packets/second:45000
Avg. Rx packets/second:45000
Avg. Tx delay (milliseconds):2
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/bs-dc-sim --nbBS=10 -duration=10 --statistic-start=3 --data-rate=1.44Mbps"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
109
'build' finished successfully (1.265s)
1
0
0
0ns
2
90000
89810
2079999ns
3
90000
90000
2079999ns
4
90000
90000
2079999ns
5
90000
90000
2079999ns
6
90000
90000
2079999ns
7
90000
90000
2079999ns
8
90000
90000
2079999ns
9
90000
90000
2079999ns
10
90000
90000
2079999ns
Avg. Tx packets/second:90000
Avg. Rx packets/second:90000
Avg. Tx delay (milliseconds):2
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/bs-dc-sim --nbBS=15 -duration=10 --statistic-start=3 --data-rate=1.44Mbps"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.264s)
1
0
0
0ns
2
135000
134715
2079999ns
3
135000
135000
2079999ns
4
135000
135000
2079999ns
5
135000
135000
2079999ns
6
135000
135000
2079999ns
7
135000
135000
2079999ns
8
135000
135000
2079999ns
9
135000
135000
2079999ns
10
135000
135000
2079999ns
Avg. Tx packets/second:135000
Avg. Rx packets/second:135000
Avg. Tx delay (milliseconds):2
[phuynh@scad ns-3.9]$ ./waf --run "examples/sg-sim/bs-dc-sim --nbBS=20 -duration=10 --statistic-start=3 --data-rate=1.44Mbps"
Waf: Entering directory `/root/ns-allinone-3.9/ns-3.9/build'
Waf: Leaving directory `/root/ns-allinone-3.9/ns-3.9/build'
'build' finished successfully (1.271s)
1
0
0
0ns
2
180000
179620
2079999ns
3
180000
180000
2079999ns
4
180000
180000
2079999ns
5
180000
180000
2079999ns
110
6
180000
180000
2079999ns
7
180000
180000
2079999ns
8
180000
180000
2079999ns
9
180000
180000
2079999ns
10
180000
180000
2079999ns
Avg. Tx packets/second:180000
Avg. Rx packets/second:180000
Avg. Tx delay (milliseconds):2