Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

28. Dual-Trigger Handover Algorithm for WiMAX Technology
Prepared by Nabil Al-Rousan (M. A. Sc. student)

IEEE 802.16e is a Worldwide Interoperability Microwave Access (WiMAX) standard that supports mobility. Handover is one of the most important factors that affect the performance of a WiMAX network. Various handover schemes have been proposed and implemented. In this paper, we propose Dual-Trigger Handover (DTHO) algorithm for WiMAX networks. The proposed handover algorithm depends on the computation of signal to noise ratio (SNR) received at the Mobile Station (MS) from various Base Stations (BSs). Relying on SNR measurements and free capacity measurements of the serving BS and the target BS improves the accuracy of handover decisions. The handover is not triggered by the MS node or the BS node individually. Instead, it is a combined decision between the two nodes. The proposed algorithm is implemented in both MS and BS nodes. We implemented the proposed algorithm using OPNET Modeler version 14 running on Windows operating system. The algorithm was simulated using multiple scenarios with various channel parameters.

27. Comparison of WiMAX and ADSL Performance when Streaming Audio and Video Content
Prepared by Tanjila Farah and Rajvir Gill (M. A. Sc. students)

The IEEE 802.16 Worldwide Interoperability for Microwave Access (WiMAX) standard is widely used for fixed and mobile Internet access. WiMAX provides maximum data rate of 75 Mbps and high-speed Internet access to a wide range of devices used by clients over the last mile. Asymmetric Digital Subscriber Line (ADSL) is widely used to provide guaranteed service.

In this paper, we compare performance of WiMAX and ADSL by streaming audio and video contents. File Transfer Protocol (FTP), Hyper Text Transfer Protocol (HTTP), and electronic mail have also been used for the comparison. We used OPNET Modeler versions 15.0 and 16.0 to evaluate packet loss, delay, delay jitter, and throughput with various design parameters to determine whether WiMAX exhibits performance comparable to ADSL.

26. Performance Analysis of Routing Protocols for Wireless Ad-Hoc Networks
Prepared by Sukhchandan Lally (M. A. Sc. student)

Wireless ad-hoc networks are decentralized wireless networks that do not rely on an infrastructure, such as base stations or access points. Routing protocols in ad-hoc networks specify communication between routers and enable them to select routes between a source and a destination. The choice of the routes is performed by routing algorithms. In this paper, we use OPNET Modeler version 16.0 A to simulate three routing protocols for wireless ad-hoc networks in several Transmission Control Protocol (TCP) and User Datagram Protocol (UDP) scenarios. We analyze route discovery time, end-to-end delay, download response time, and routing traffic overhead in static, less dynamic, and highly dynamic mobility scenarios. Simulation results indicate that Ad-Hoc On-Demand Distance Vector (AODV) protocol is the most flexible when compared to Dynamic Source Routing (DSR) and Optimized Link State Routing (OLSR) protocols in the case of movement. OLSR is the only protocol that meets the end-to-end delay requirements of less than 20 ms.

25. Simulation and Performance Evaluation of WiFi and WiMAX using OPNET
Prepared by Ravinder Paul and Sukhchandan Lally (M. A. Sc. students)

Wireless Fidelity (WiFi) network is based on the IEEE 802.11 standard. Worldwide Interoperability for Microwave Access (WiMAX), based on IEEE 802.16, is a standard with similar principles. The main advantage of WiMAX over WiFi is that it covers larger areas and has higher data rates. WiMAX network operators provide WiMAX subscriber units that enable connection to the metropolitan WiMAX network while WiFi units are used for connecting local devices within homes or businesses. In this paper, we use OPNET Modeler to simulate and compare WiFi and WiMAX in a small area network and compare their performance in terms of mobility. Simulation results indicate that WiMAX may carry larger load and has better throughput.

24. Performance Analysis of RIP, EIGRP, and OSPF using OPNET
Prepared by Don Xu (Ph.D. student)

Routing protocols are key elements of modern communication networks. Currently deployed dynamic routing protocols that are used to propagate network topology information to the neighboring routers are Routing Information Protocol (RIP), Enhanced Interior Gateway Routing Protocol (EIGRP), and the Open Shortest Path First (OSPF) protocol. The choice of the right routing protocol depends on a number of parameters. In this paper, we use OPNET Modeler to analyze the performance of RIP, EIGRP, and the OSPF protocols, which are commonly deployed in Internet Protocol (IP) networks. We designed various simulation scenarios to compare their performance.

23. Performance Evaluation of Transport Protocols for Internet-Based Teleoperation Systems
Prepared by Jae-young Lee (Ph.D. student)

An Internet-based teleoperation system is an interactive application where a human user transmits movement data of a robotic device while simultaneously receiving reflected force data from a remote teleoperator. Performance of such real-time applications is highly sensitive to the Internet delay and data loss. In this paper, we describe the efficient transport protocol (ETP) designed for Internet-based teleoperation systems and demonstrate that it reduces the round trip time (RTT) between a human user and a remote teleoperator. We use OPNET Modeler to simulate the ETP protocol and compare its performance with the transport control protocol (TCP) and the user datagram protocol (UDP) by measuring end-to-end delays in various simulation scenarios. We observe that with UDP as a transport protocol, ETP reduces the end-to-end delay by introducing an inter-packet gap (IPG).

22. Streaming Video and Audio Content over Mobile WiMAX Networks
Prepared by Will Hrudey (M. Eng. student)

WiMAX (Worldwide Interoperability for Microwave Access) embodies the IEEE 802.16 family of standards that provision wireless broadband access. With the IEEE 802.16e-2005 mobility amendment, WiMAX promises to address the ever-increasing demand of mobile high-speed wireless data in fourth generation (4G) networks. WiMAX market studies continue to project increased subscriber growth rates and planned carrier trials worldwide. With these projected growth rates, in order to understand if WiMAX is a formidable player in 4th generation mobile systems, it is desirable to quantify performance using video-rich emerging services to sufficiently load and stress the network to exploit the potential bandwidth, delay and mobility limitations.

The goal of this project is to enhance the existing OPNET simulation model "Streaming video content over WiMAX and ADSL access networks" (Model IDs: 977, 899, and 896) to simulate video content representative of IPTV and other video-rich emerging services to adequately load and analyze the Mobile WiMAX technology.

21. Streaming Video Content Over IEEE 802.16/WiMAX Broadband Access
Prepared by Will Hrudey (M. Eng. student)

Worldwide Interoperability for Microwave Access (WiMAX) embodies the IEEE 802.16 family of standards that provide wireless broadband access to residential and commercial Internet subscribers. While other WiMAX applications exist, there is an increasing trend to employ WiMAX for last-mile Internet access to circumvent the high deployments costs and local loop distance limitations associated with wired Asymmetric Digital Subscriber Line (ADSL) connections.

We use the OPNET Modeler to simulate bandwidth intensive, delay sensitive, video traffic representative of Internet Protocol Television (IPTV) and other video-rich applications over WiMAX and ADSL. These video streams are typically encoded using MPEG-x codecs. Although marginally loss-tolerant, performance of these streams is inherently a function of available bandwidth, buffering, and delay characteristics of the underlying network. Hence, in this paper, we examine four performance factors while streaming two hours of video content to client subscribers to determine whether WiMAX can deliver access network performance comparable to ADSL for video applications.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

20. Performance Evaluation of TCP Tahoe, Reno, Reno with SACK, and NewReno Using OPNET Modeler
Prepared by Laxmi Subedi and Mohamadreza Najiminaini (M. A. Sc. students)

Over the past two decades, numerous Transmission Control Protocol (TCP) congestion control algorithms have been proposed for deployed wireless and wired networks. The main goal of these algorithms was to successfully manage congestion, reliably handle loss, and minimize transmission errors. In this paper, we use OPNET Modeler to simulate four TCP versions (Tahoe, Reno, Reno with Selective Acknowledgment (SACK), and NewReno) in several wireless and wired networks. We analyze congestion window maintenance and recovery process for each TCP algorithm by observing congestion window size, file download response time, throughput, and goodput. We observe that in wireless networks with signal attenuation, fading, and multipath, TCP Reno has larger congestion window size, shorter file download response time, and higher throughput and goodput than the remaining three TCP algorithms. In case of wired networks, TCP Reno with SACK

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

19. OPNET Model of TCP with Adaptive Delay and Loss Response for Broadband GEO Satellite Networks
Prepared by Modupe Omueti (M. A. Sc. student)

Transmission control protocol (TCP) provides reliable transport services for Internet applications. Broadband geostationary earth orbit (GEO) satellite networks play an important role in providing Internet access and network connectivity. They are characterized by long propagation delays and high bit error rates, which negatively affect TCP performance. We proposed a modification of TCP named TCP with adaptive delay and loss response algorithm (TCP-ADaLR) that improves TCP performance in broadband GEO satellite networks.

In this paper, we describe the TCP-ADaLR algorithm and its OPNET implementation. We evaluate and compare the performance of TCP-ADaLR, TCP SACK, and TCP NewReno, with and without delayed acknowledgment. To verify the algorithm implementation, we first consider simulation scenarios with an ideal lossless satellite link. We then consider simulation scenarios with congestion losses and show that TCP-ADaLR exhibits comparable performance to TCP SACK and TCP NewReno. In the presence of losses due to satellite link errors, TCP-ADaLR outperforms TCP SACK and TCP NewReno. In the presence of both congestion and error losses, TCP-ADaLR improves TCP goodput and throughput. In all simulation scenarios, TCP-ADaLR outperforms TCP SACK and TCP NewReno in terms of satellite link throughput and the user-perceived latency of HTTP web and FTP file download applications. TCP-ADaLR is fair to competing connections, friendly to TCP NewReno, and maintains the end-to-end semantics of TCP.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

18. M-TCP+: Using Disconnection Feedback to Improve Performance of TCP in Wired/Wireless Networks
Prepared by Modupe Omueti (M. A. Sc. student)

In this paper, we propose the M-TCP+ algorithm for heterogeneous wired/wireless networks. The algorithm is a modification of M-TCP that was proposed for deployment in mobile cellular networks. It is recommended that Internet hosts enable the delayed acknowledgement (delayed ACK) option to maximize network bandwidth by reducing the number of ACKs sent to a TCP sender by a TCP receiver. The M-TCP+ algorithm performs best when the TCP delayed ACK option is enabled. The algorithm relies on feedback sent from a wireless host in anticipation of disconnections. We compare the performance of the M-TCP+ algorithm with the performance of M-TCP, TCP NewReno, and TCP SACK in both the absence and the presence of disconnections for a file transfer protocol (download) application. We also simulate network scenarios with traffic congestion. The M-TCP+ algorithm performance is evaluated in terms of file download response time, goodput, and retransmission ratio with and without the delayed ACK option. In scenarios without disconnections, the M-TCP+ algorithm does not introduce significant processing delay. Furthermore, in scenarios with disconnections, the M-TCP+ algorithm shows 2%-15% performance improvement.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

17. General Packet Radio Service OPNET Model
Prepared by Renju Narayanan (M. A. Sc. student)

In this paper, we describe a General Packet Radio Service (GPRS) OPNET simulation model and the implementation of the Radio Link Control/Medium Access Control (RLC/MAC) and the Base Station Subsystem GPRS protocol (BSSGP). The RLC/MAC and BSSGP protocols are added to an existing GPRS OPNET model. We have enhanced the existing model by implementing unacknowledged mode of RLC and two phase access mechanisms. The implementation of BSSGP enables the exchange of radio-related and data messages from Base Station Subsystem (BSS) to Serving GPRS Support Node (SGSN). We have verified the effect of the new implementation on the end-to-end delay and cell update mechanism by performing OPNET simulations. The enhanced model was tested using a network with 17 mobile stations.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

16. Effect of Cell Update on Performance of General Packet Radio Service
Prepared by Modupe Omueti and Renju Naryanan (M. A. Sc. students)

In this paper, we describe simulation and evaluate the effect of cell update on General Packet Radio Service (GPRS). GPRS supports packet-switched services in cellular networks. We use a GPRS OPNET simulation model that implemented GPRS-specific communication protocols. The developed OPNET model supports two QoS profiles based on the mean throughput class. We validate the GPRS implementation based on the observed link throughput between base transceiver stations and the base station controller. We evaluate the effect of cell update on the end-to-end delay, time to process signaling messages, and throughput. Simulation scenarios with and without cell updates are employed to illustrate that signaling processing time and the delay as perceived by the user increase with cell update.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

15. Enhanced General Packet Radio Service OPNET Model
Prepared by Renju Narayanan (M. A. Sc. student)

In this paper, we describe the enhancements made to an existing General Packet Radio Service (GPRS) network OPNET model. These enhancements to the existing model are the implementations of the Logical Link Control (LLC) layer, the Base Station Subsystem (BSS), and the cell update procedure. We first present an overview of the GPRS network, the LLC layer, and the existing OPNET model. We then describe the implementation of the LLC layer, the BSS consisting of a Base Station (BS) and a Base Station Controller (BSC), and the autonomous cell reselection procedure performed by the mobile station. Four simulation scenarios were used to verify the accuracy of the OPNET implementation. We conclude by suggesting possible further enhancements to the GPRS OPNET model.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

14. OPNET Implementation of the Megaco/H.248 Protocol: Multi-Call and Multi-Connection Scenarios
Prepared by Edlic Yu and Edwood Yu (M. Eng. student)

In this paper, we describe the OPNET implementation of MEGACO/H.248 signaling protocol. The OPNET model allows for multi-call and multi-connection scenarios, where any number of Media Gateways (MGs) can simultaneously connect to the Media Gateway Controller (MGC). This design is important for simulations of simultaneous voice conversations. In our simulation scenario, multiple MGs can connect to one MGC via a router. We simulated the call-establishment, call-waiting, and call-release scenarios by employing a complete set of MEGACO/H.248 signaling commands. We also simulated voice transmission using packets encoded with the Real-Time Transport Protocol (RTP).

Professor: Stephen Hardy and Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

13. Compressed Real-Time Transport Protocol (cRTP)
Prepared by Kevin Ko (M. Eng. student)

The cRTP process model performs RTP/UDP/IP header compression according to a subset of the cRTP (IETF RFC 2508) rules. RTP packets are received from an upper layer application in an IP format (ip_v4_pkt, custom). The IP payload is UDP (udp_v2_pkt, custom), with its own payload of RTP format (rtp_pkt, custom). The header is compressed when possible and a session context is kept. The cRTP algorithm includes only basic operation to handle real-time traffic. They are sufficient for simulation and testing purposes. The cRTP model has no promoted parameters.

The cRTP process model does not include several functionalities described in RFC 2508. They are necessary to complete the model:

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

12. Enhancements and performance evaluation of wireless local area networks
Prepared by James Song (M. A. Sc. student)

Unlike wired networks that can provide large bandwidth, the bandwidth of wireless local area networks (WLANs) is rather limited because they rely on an inexpensive, but error prone, physical medium (air). Hence, it is important to improve their loss performance.

In this paper, we investigate several methods for improving the performance of WLANs. We survey the current research literature dealing with improving performance on various wireless network layers. We describe OPNET implementations of three approaches: tuning the physical layer related parameters, tuning the IEEE 802.11 parameters, and using an enhanced link layer (media access control) protocol. Finally, we describe several simulation scenarios and present simulation results that demonstrate the effectiveness of the three approaches.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

11. Cellular Digital Packet Data (CDPD) MAC layer model
Prepared by Savio Lau and Eric Keung (B. A. Sc. students)

The Cellular Digital Packet Data (CDPD) Medium Access Control (MAC) layer network model consist of multiple Mobile End Stations (M-ES) and a single Mobile Data Base Station (MDBS). The model contains several enhancements compared to the previous CDPD MAC model.

The model was used to investigate the impact of various traffic sources (constant bit rate, Poisson, non-heavy-tailed bursty, heavy-tailed) on the M-ES buffer occupancy.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

10. OPNET implementation of the Mobile Application Part protocol
Prepared by Vladimir Vukadinovic (M. A. Sc. student)

This paper describes OPNET implementation of the Mobile Application Part (MAP) protocol within the General Packet Radio Service (GPRS) model. MAP represents an application layer protocol residing on top of the Signaling System 7 (SS7) protocol stack. In GPRS networks, MAP protocol supports signaling exchanges with Home Location Register (HLR) and Equipment Identity Register (EIR). We begin with a brief introduction to SS7 protocol stack and GPRS architecture. We then describe MAP features related to

GPRS and modifications that we implemented in the GPRS OPNET model to provide signaling capabilities for communication between HLR and the Serving GPRS Support Node (SGSN). We provide implementation details and evaluate the impact of the MAP protocol overhead on network response time.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

9. OPNET implementation of endpoint admission control algorithm
Prepared by Karen Wu and Yan Ma (M. A. Sc. students)

The history of telephone networks indicates a trend of employing sophisticated switches and dumb hosts. Hence, it is often expensive and difficult to replace switches and add new applications. In contrast, Internet employs simple routers and sophisticated applications in hosts. New applications can be added without requiring infrastructural changes. In order to keep routers simple, there is a growing need for new mechanisms employing end-hosts that provide admission control.

In this paper, we describe OPNET implementation of two endpoint admission control (EAC) algorithms. The OPNET EAC model captures basic EAC procedures for network probing and data transmission. We discuss several architectural issues and implementation details. We then describe the packet format and the node and process models of the end-hosts and routers. We also present simulation scenarios based on various traffic sources, traffic loads, probing times, and number of hops.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

8. Performance evaluation of M-TCP over wireless links with periodic disconnections
Prepared by Wan G. Zeng, Meihua Zhan, and Zhiwen Li (M. A. Sc. students)

Since 1988, when the congestion control functions were first introduced, the performance of TCP has been greatly improved. However, TCP still suffers large performance degradation over wireless links due to characteristics specific to wireless environment that affect the behavior of TCP's congestion control and avoidance mechanisms. These mechanisms were designed and optimized for traditional wireline networks where packet losses are predominantly due to network congestion. In wireless networks, packet losses are mainly caused by the high bit error rate (BER) and hand-offs. Since TCP cannot differentiate packet losses caused by congestion from losses introduced by the wireless links, its performance degrades. M-TCP is a solution proposed to address the problems of TCP over wireless links with periodic disconnections.

In this paper, we first give a brief introduction to M-TCP protocol. We then describe the OPNET implementation of M-TCP and the simulation scenarios in a mixed wireline/wireless environment. We give performance comparisons between M-TCP and TCP with and without presence of frequent disconnections in wireless links. Simulation results indicate that in the presence of frequent disconnections M-TCP outperforms TCP in terms of maintaining congestion window size, goodput, and sender size retransmission timer.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

7. OPNET implementation of the Megaco/H.248 protocol
Prepared by Shufang Wu, Mahmood Riyadh, and Riadul Mannan (M. A. Sc. students)

The best-known media gateway control protocols are currently Media Gateway Control Protocol (MGCP) and Megaco/H.248. MGCP, described as informational RFC 2705, has been widely deployed. Megaco/H.248 has evolved from MGCP. It is a single unified protocol that resulted from the cooperation between the Internet Engineering Task Force (IETF) and the International Telecommunication Union (ITU). Megaco/H.248 is expected to win wide industry acceptance as the official standard.

In this paper, we describe OPNET implementation of Megaco/H.248. Our simulation scenario employs one Media Gateway Control (MGC) and two Media Gateways (MGs) connected to a local hub by point-to-point duplex links. We have implemented five Megaco/H.248 commands in order to simulate three basic call flows and demonstrate Megaco/H.248 functionality. Data traffic between two MGs is simulated using the Real-Time Transport Protocol (RTP). We describe the OPNET implementation details of the three basic call flows: Successful MG Registration Procedure, Successful Call Setup Procedure, and Successful Call Release Procedure. We also present simulation results and discuss possible future enhancements of our implementation.

6. Simulation of General Packet Radio Service network
Prepared by Ricky Ng (M. Eng. student)

In this paper, we describe an OPNET model of a General Packet Radio Service (GPRS) network. The model captures the signaling and transmission behavior of the GPRS network. We first introduce a GPRS network and describe the signaling and transmission procedures that will be modeled. In order to point out the simplifications made in the OPNET model, we then address the differences between the GPRS OPNET model and the standard. We also describe the implementation of each model component: node model, packet format, process model, and state variables corresponding to each component. Furthermore, we illustrate the use of OPNET to abstract the signaling and transmission behavior of the GPRS network.

Discussion of the simulation results follows the implementation details. The GPRS network consists of sources with two distinct classes of Quality of Service (QoS), connected to a sink. We consider two simulation scenarios. The first scenario is used to confirm that the GPRS network model allows a Mobile Station (MS) that has subscribed to a GPRS service to access the network, to set up a data session, and to transfer uplink user data. Simulation results of the packet end-to-end delay also confirm that the two classes of QoS are correctly implemented. In the second simulation scenario, we capture various statistics to illustrate the model's performance. We conclude by addressing how the model can be deployed as a performance assessment tool in a genuine GPRS network.

5. Performance evaluation of TCP over WLAN 802.11 with the Snoop performance enhancing proxy
Prepared by Chi Ho Ng and Jack Chow (M. Eng. students)

The growing popularity of wireless devices used to access the Internet and an increasing use of the TCP/IP protocol suites, indicate that in the near future TCP protocol will be frequently used over the wireless links connecting wireless devices. The characteristics of wireless links are significantly different from characteristics of wired network links because data is frequently lost due to the volatile environment in which wireless links operate. TCP was originally designed for wired networks, where loss of data is assumed to be due to congestion only. This assumption leads to poor performance of TCP in a wireless environment. Hence, various mechanisms were proposed to improve TCP performance over wireless links. One such mechanism is a performance enhancing proxy, such as the Snoop protocol.

In this paper, we study the effect of the Snoop protocol on the performance of TCP over wireless links. We implemented and simulated Snoop protocol in OPNET wireless LAN (WLAN) devices. We measured the performance improvement by comparing the performance of an FTP session with and without the Snoop protocol. We found that the Snoop protocol significantly improves the performance of TCP (~ 68 times under 30% packet error rate). We show that this improvement is achieved by preventing the TCP layer from reducing the congestion window size.

4. Simulation of congestion control algorithms using OPNET
Prepared by Nazy Alborz (M. A. Sc. student)

We use OPNET to model congestion avoidance algorithms, such as policing and scheduling schemes in packet switched networks, in order to analyze network performance parameters (cell loss and delay). We vary control parameters specific to each algorithm and we use genuine traffic traces from the existing networks for packet generation.

4.a VirtualClock algorithm:
We are currently implementing the "VirtualClock" scheduling scheme in OPNET. VirtualClock is a traffic control mechanism for packet switched networks. It is used for scheduling the packet queues and it is implemented inside the network routers.

A process model named ip_output_iface_pr.m, avilable in OPNET 7.0, models various queuing mechanisms such as FIFO, WFQ, or priority queuing. This process model is a child process that is called by the IP layer process model of each IP router (advanced) node. The ip_output_iface_pr.m process model consists of two functions: a classifier function responsible for enqueuing the packets, and a scheduler function.

The classifier function builds queues for the packets according to their selected scheduling algorithm and assigns each packet to a queue. The scheduler is in charge of prioritizing the packets in each queue and in harge of scheduling the service time for them.

In order to implement the VirtualClock algorithm the classifier has to classify the packets in a per flow basis. It means that packets of each flow are assigned a separate queue and the scheduling is done according to the virtual clock tick of each flow.

The statistics that can be gathered from the IP router containing VirtualClock algorithm are:

Two of them, the queuing delay and traffic dropped, give us the possibility for analyzing the VirtualClock algorithm functionality under different types of traffic models like Poisson, self-similar traffic, and genuine traces gathered form the real networks.

4.b Leaky bucket algorithm:
Leaky bucket is one of the policing algorithms we simulated using OPNET. This algorithm is designed for ATM networks. It is implemented at the user-network interface (UNI) to monitor the cell rate of the traffic source so that the source does not exceed the negotiated mean cell rate. The maximum bit cell rate of the source can also be monitored by using in sequence two leaky buckets with distinct parameters.

We have modeled the leaky bucket algorithm by creating a process model consisting of several states. Arrival of each packet takes the algorithm to the "ARRIVAL" state, in which the difference between actual and negotiated cell rate is examined. If the cell rate remains within a specific limit, the cell will be passed to the network by entering PK_SERVED state and it will be served according to the guaranteed performance. Otherwise, the cell will be dropped by entering PK_DROPPED state and will never be served. Users can define the size of the cell burst (bucket size) and the negotiated cell rate (leaking rate) as input parameters to the process model for every single source implementing the algorithm in the network.

Professor: Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

3. OPNET implementation of IPv6 type of service over ATM network
Prepared by Ricky Ng and Danny Yip

We investigate the application and the mapping of the IPv6 Type of Service (ToS) onto various Asynchronous Transfer Mode (ATM) Quality of Service (QoS) classes. Our goal is to employ different ToS values offered by IPv6 and distinct service categories and QoS classes offered by ATM network so that IPv6 packets may be switched and delivered to their destination based on the specified ToS header fields. Such an IPv6-over-ATM network may then deliver IPv6 packets according to their specified ToS, and packets with distinct ToS will experience different network end-to-end delays.

In our implementation, we used three existing OPNET models: atm_uni_src, atm_crossconnect, and atm_uni_dest. They are used to model an ATM network that allows the ATM sources (atm_uni_src) to generate raw packets employing any of the five ATM service categories. These packets, encapsulated by the ATM Adaptation Layer 5 (AAL5), are segmented into ATM cells (atm_crossconnect) and are sent to an ATM switch that routes them to the destination (atm_uni_dest) where they are re-assembled into AAL5 packets. We implemented an IPv6-over-ATM interface module that determines the ATM Switched Virtual Circuit (SVC) that each IPv6 packet should occupy when delivering packets to their destination. The simulated network consists of a host that uses the new module. Simulation result revealed that IPv6 packets with higher priority (specified by the ToS field) experience smaller packet end-to-end delay. These results indicate the feasibility of using the ToS header byte in IPv6 packets in meeting various ATM QoS requirements.

Professor: Tiko Kameda
Department: School of Computing Science

2. Differential service for Internet
Prepared by George Fang (M. A. Sc. student)

I am working on simulating the design and implementation of a traffic condition block for an internet router using DiffServ architecture. Logically the main part of a traffic condition block consists of classifying, metering and transmission. These parts interact with each other and with other parts of a router. To have a fast traffic condition block, we need to use multiple processes on multiple cpus to implement these essential parts. Understanding the interaction and serialization of these parts is important for a fast algorithm. OPNET offers a rather convenient and easy way to do so as currently it is very hard to find and use a computer with multiple cpus running true multi-tasking operating systems. To carry out the simulation using opnet, we first have to gather reliable raw data for individual parts using regular simulation technique and then feed the data into opnet. Using such an approach we also hope to have a better idea of how to divide up the traffic condition block into various parts. In the end, we shall also find out the optimal number of processes to use to implement the main conditioning block.

The data obtained will be further used to investigate the complex interaction of many routers within a network. However from the current experience with the simulation, we feel the whole thing may be too complex and too slow to be studied fully.

Professors: Stephen Hardy and Ljiljana Trajkovic
Department: School of Engineering Science
Laboratory: Communication Networks Laboratory

1. Analysis and simulation of wireless data network traffic
Prepared by Michael Jiang (M. A. Sc. student)

We investigate traffic patterns in wireless data networks. Preliminary results demonstrate that wireless data traffic exhibits long-range dependent behavior, and, therefore, is statistically different from traffic generated by traditional traffic models. To evaluate the performance of wireless Cellular Digital Packet Data (CDPD) networks, we use the OPNET (Mil. 3) simulator using genuine traffic traces from the Telus (formerly BCTel) wireless network. Our simulation results indicate that genuine traffic produces longer queues and thus requires larger buffers in the network's switching elements.

OPNET Technologies, Inc. is a leading provider of solutions for managing networks and applications. OPNET's best-in-class solutions address application troubleshooting, application monitoring, network monitoring, network configuration management, network planning , and network simulation. OPNET’s solutions have been operationally proven in thousands of customer environments worldwide, including corporate and government enterprises, government and defense agencies, network service providers, and network equipment manufacturers. For more information about OPNET and its products, visit www.opnet.com.