image001.gif

 

 


OPNET Technologies
7255 Woodmont Avenue
Bethesda, MD 20814

Tel: 240-497-3000
Fax: 240-497-3001
E-mail: university@opnet.com
Web: www.opnet.com

2008 OPNET Technologies, Inc.
OPNET and OPNET Technologies Inc.
are registered trademarks of OPNET.
All trademarks are the property of their
respective owners and are used herein
for identification purposes only.


University: McGill University
Name of sponsoring Professor: Tho Le-Ngoc
Department: Department of Electrical and Computer Engineering

 

Research:

 

Round-robin scheduling (RR)

 

Round-robin scheduling (RR) is one of the simplest packet scheduling algorithms. RR assigns time slices to each queue in equal portions and in order, handling all queues without priority. RR is both simple and easy to implement, and starvation-free.

 

Priority queuing (PQ)

 

Priority queuing supports some number of queues, usually from high to low. Queues are serviced in strict order of queue priority, so the high queue always is serviced first, then the next-lower priority and so on. Lower-priority queues are scheduled only when there is no backlogged packet in any higher-priority queue.

 

Weighted Fair Queuing (WFQ)

 

Weighted Fair Queuing (WFQ) is a packet scheduling technique allowing guaranteed bandwidth services. The purpose of WFQ is to let several sessions share the same link. WFQ is an approximation of Generalized Processor Sharing (GPS) which, as the name suggest, is a generalization of Processor Sharing (PS). In PS each session has a separate FIFO queue. At any given time the N active sessions (the ones with non-empty queues) are serviced simultaneously, each at a rate of 1/N:th of the link speed. Contrary to PS, GPS allows different sessions to have different service shares. GPS have several nice properties. Since each session has its own queue, an ill-behaved session (who is sending a lot of data) will only punish itself and not other sessions. Further, GPS allows sessions to have different guaranteed bandwidths allocated to them.

 

The performance of WFQ (using the WFQ model we have built) is studied in different network and traffic scenario. We are looking for ways to accurately simulate the effect of variable service rate and its effect on the queuing performance.

 

Multi-time-scale traffic modeling

 

Multi-time-scale modeling of traffic using multiplicative models is investigated. Specifically, the use of Multi-scale traffic models for modeling both Long Range Dependent (LRD) and Short Range Dependent (SRD) traffic is studied. A Multi-Scale OPNET traffic model is constructed. Traffic generated form the model will be used as an input to our existing Weighted Fair Queue and Priority Queue models to evaluate the performance of these scheduling approaches.

 

Trace-file-based video traffic model

 

A video traffic generator which is based on an actual trace file is constructed. The trace file records all packets (generation time, packet length, packet type, …) of long MPEG-4 and H.263 encoded videos. The videos are of typically 60 minutes length each. This video traffic model will be used as an input to our existing Weighted Fair Queue and Priority Queue models to evaluate the performance of these scheduling approaches.

 

Rayleigh Fading Channel Model

 

This process model is used to simulate the Rayleigh fading channel. Rayleigh fading channel is a typical assumption for wireless communication systems and it is widely used. The channel is assumed to be block fading with each block being fixed (could be assigned a different value). The packet service time is proportional to the channel gain at each given block of time. The Rayleigh fading process model is used along with the queuing architecture models to obtain their random service rate.

 

ON-OFF traffic model

 

Telecommunication traffic is very bursty, or equivalently, the ratio of the peak to average data rate is quite high. To handle the bursty traffic the network must be sized to handle the peak load. Most networks today do not attempt to carry the absolute offered peak bursty traffic, but provide some degree of buffering to smooth these peaks out. As a result, bursty traffic models are an essential component of performance analyses of telecommunication networks. Good models should be simple, accurate and applicable in both mathematical analysis and computer simulations. ON-OFF model is widely used to represent bursty traffic that is generated from one source with alternating series of “ON” and “OFF” periods. Both the ON and OFF periods were distributed by different (Poisson, Lognormal, Weibull, etc.) distributions.

 

Teaching:

 

ECSE 474: Design Project 1


Electrical Engineering: A laboratory design project undertaken with close supervision by a staff member. The project consists of defining an engineering problem, reviewing relevant background and literature, and seeking the solution through numerical simulation and/or experimental investigation. A literature review, written project proposal, and seminar presentation are required.


ECSE 475: Design Project 2

 

Electrical Engineering: A laboratory design project undertaken with close supervision by a staff member. A continuation of ECSE 474 Design Project 1. The work consists of carrying out the project plan developed in ECSE 474 Design Project 1 producing a report summarizing the results, and a seminar presentation.


ECSE 498: Honours Thesis 1

 

Electrical Engineering: A research project undertaken with close supervision by a staff member. The work consists of defining an engineering problem, reviewing the associated literature, and seeking the solution through experimental investigation. A literature review and a written thesis proposal are required along with a seminar presentation at end of term.

 

ECSE 499: Honours Thesis 2

 

Electrical Engineering: A research project undertaken with close supervision by a staff member. A continuation of ECSE 498. The work consists of carrying out the research plan developed in ECSE 498 along with a seminar presentation at end of term.


 

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, capacity management, 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.