ATM Network Research:
In conjunction with the University of Saskatchewan, we are conducting simulation
studies of ATM networks to deal with traffic modeling and applying
our traffic models to benchmark existing and novel traffic admission
and policing schemes.
We are attempting to quantify the ``equivalent bandwidth'' for a variety
of traffic patterns arriving at an ATM switch with the intention of using
this quantification for making ``educated'' call admission/rejection decisions.
Multimedia applications, such as distance learning, sales training sessions, and video conferencing for large groups of users over a wide area network, need an efficient method of connecting the members for these specific groups. The routing mechanism of such group connections is called multicasting. We are involved in characterizing different group communication demands and developing fast connection algorithms which allocate the necessary resources along the routes allowing group members to communicate interactively.
We are looking at the problem of finding physical and logical topologies of ATM networks. A prototype tool has been developed to find a low-cost, physical design for interconnecting ATM switches in a local area in order to provide quality of service (QoS) guarantees for projected traffic. This methodology guarantees QoS requirements with respect to loss and provides routing information by creating a virtual path between each source and destination. We are also studying the problem of finding virtual path layouts in large ATM networks (this project is financed by the Canadian Institute of Telecommunication Research).
Non-ATM Networks:
This part of our research deals with novel deflection schemes for high-speed
datagram networks.
Our plans include the investigation of asynchronous deflection
schemes applicable to networks with a higher connectivity than the
2x2 switch architecture of Manhattan-street networks (MSN).
We have proposed a deflection scheme applicable to 4x4 torus grids,
which is asynchronous and imposes a limit on the maximum number of hops
traveled by a packet on its way to the destination.
Multimedia Applications:
We been developing a software package for distance education.
So far this work resulted in a new video encoding algorithm,
dubbed spaflay, especially well suited for applications
in teleconferencing and distance education.
Also, a prototype software package, called CyberSchool, is
aimed at the development of hyper-media software for a remote school,
integrating text, images, audio, and video with the objective of delivering
an effective presentation of the course material to a remote student.
Issues such as location-transparency, doubt clarification, information
integrity, and storage economy are being addressed.
Our analyses of medical imaging networks standards have led to significant changes in the former ACR-NEMA standards. We are now starting to investigate two practical problems: 1) hospital LANS or MANS with an ATM backbone network for fast transmission and interactive use of large images; and 2) an investigation and assessment of how, when, and why storage devices and databases may be the bottleneck for fast medical networks.
Reserving resources in advance can solve congestion problems in networks. Many multimedia applications such as video conferencing and video distribution are suited for advanced scheduling. Scheduling algorithms have been found for video conferencing applications. We are now investigating the use of advanced reservation for video distribution services.
Tools:
Network performance, including delay, delay variance, bandwidth utilization,
fairness, is studied by analytical techniques and simulation.
Simulation is carried out with the aid of locally developed simulation
packages
LANSF
and
SMURPH.
We plan to adapt SMURPH to modeling mobile networks at the bit level.
As part of a joint effort with the
University of Calgary
and the
University of Saskatchewan
(the TeleSim project financed by CANARIE and
industrial sponsors), we have developed a high-fidelity, high-performance
simulator for ATM networks.
You are visitor to the Communications Networks Research Group home page.