[IEEE Kite] [CCECE02 Logo]
May 12 to 15, 2002
Hotel Fort Garry, Winnipeg, Manitoba, Canada

[Call for Papers]
[IEEE Canada]
[Contact Us]

Advances in Telecommunications for the 21st Century:
Selected Issues and Approaches

Professors Bob McLeod, Muthucumaru Maheswaran, and Ekram Hossain
Department of Electrical & Computer Engineering
University of Manitoba
Winnipeg, MB R3T 5V6
Phone: (204) 474-9603
Fax: (204) 261-4639

12 May 2002

1:00 to 5:00 PM

University of Manitoba
Engineering Bldg.
Fort Garry Campus
Winnipeg, MB R3T 5V6

  1. Part A: TCP/IP and Transport Layer Enhancements
  2. Part B: Next-Generation Middleware Platforms
  3. Part C: MAC Design for Mobile Pervasive Ad hoc Networking

PART A: TCP/IP and Transport Layer Enhancements

(Professor Bob McLeod)
It is hard to describe network computing, the Internet, wireless, microelectronics, or fiber optics as "emerging" areas in Information and Communication Technology. It would be more accurate to describe these as periodically having coming out parties. In spite of the impressive gains these areas have seen in terms of underlying technology, relatively little has changed regarding the protocols for moving data. For example, the 802 LAN standard had its origin in February 1980 (hence denoted 802), with the basic protocol such as 802.3 not changing much but undergoing significant performance advances due to the technology of the day. Low level protocols such as 802.3, or the 802.11, are based on common sense and a "what else would you do mentality", and as a consequence work well. The picture is less clear with respect to WANs where we are still using protocols that were developed over 30 years ago that don't have the same common sense pedigree. Brilliant people dreamt up the TCP/IP suite of protocols likely in a drug induced Gedanken experiment as a means of reliably moving a small amount of packet data through a highly unreliable and relatively small network. The idea presented here is that it is time to revisit some of these protocols and consider the development or invention of efficient transport level protocols as a re-emerging area of communication technology.

This section of the tutorial will review TCP/IP with emphasis on the transport layer and the impact that defacto standard layer physical layer technologies such as Ethernet has on them. A review of existing and proposed alternative transport layer protocols will be provided as well as an overview of research we are doing at the University of Manitoba.

PART B: Next-Generation Middleware Platforms

(Professor Muthucumaru Maheswaran)
The Internet has evolved into a popular communication medium mainly driven by applications such as E-mail, Net News, and World Wide Web. These applications primarily require the Internet to be a best effort data mover to enable "information exchange." However, several existing and emerging applications from areas such as business, entertainment, health, and education require services beyond the best-effort model. This is further complicated by the emergence of different networking technologies and the desire of the end-users to access services from everywhere at anytime.

To address these requirements, recent research has focused on the development of new scalable, efficient, and flexible "next generation" middleware platforms. These middleware platforms are intended to support the deployment of wide-area applications such as Content delivery networks and Video-on demand. One such next-generation middleware platform is the Grid computing system. A Grid is a virtual framework that allows resources across national and international boundaries to interoperate and share capacities in a controlled fashion to form scalable systems.

This talk introduces the concept of Grid computing and includes the various different motivations for developing Grid systems. In particular, we emphasize on the use of Grid computing systems for supporting wide-area applications by providing a resource provisioning framework. Architectures of existing Grid systems is also presented. Following the discussion of the architectures, the talk will examine the different issues in developing and deploying a large-scale middleware as the Grid system.

To illustrate some of the research issues involved in Grid systems, we discuss resource discovery, dissemination, scheduling and security issues in this context. Simple examples will be used to highlight the issues and the methods developed to address them. Time permitting, we will demonstrate some of the concepts using an advanced Grid prototype that is being developed at TRLabs and University of Manitoba.

PART C: MAC Design for Mobile Pervasive Ad hoc Networking

(Professor Ekram Hossain)
The simplicity in deployment makes wireless ad hoc networks (which do not require any infrastructure) suitable for a variety of applications such as collaborative computing, disaster recovery, battle field communication. With the proliferation of communications and computing devices such as cellphones, laptops, or PDAs, Personal Area Networking (PAN) which is an ad hoc networking-based technology, has recently gained much interest.

MAC protocols play an important role in the performance of the wireless ad hoc networks. A MAC protocol defines how each mobile unit can share the very limited wireless bandwidth resource in an efficient manner. Recent researches in this area have focused on designing MAC protocols with optimized performance metrics including throughput, delay, fairness, stability, support for multimedia and power efficiency.

In this section of the tutorial, a thorough survey of all the existing medium access control approaches is presented. The most of the complexity in MAC design for wireless ad hoc networks arises due to node mobility and link vulnerability. A series of studies on MAC design has been conducted to improve medium access performance in different aspects as identified by the different performance metrics. Tradeoffs among the different performance metrics (such as between throughput and fairness) dictate the design of a suitable MAC protocol. We compare the different proposed MAC approaches, identify their problems, and discuss the possible remedies. To this end, the interactions among the MAC and the higher layer protocols such as routing and transport layer protocols are discussed. We also shed light on some of our current research efforts in this area.


Prof. Ekram Hossain, Ph.D.
Department of Electrical & Computer Engineering
University of Manitoba
Winnipeg, MB R3T 5V6

Each workshop requires payment of fees of $30 / $50 for IEEE members and students (before March 15 / after March 15) and $50 / $80 for IEEE non-members (before March 15 / after March 15). The fee will cover workshop materials and refreshments.

[Small Logo] Your suggestions and questions are always welcome.
Copyright © Notice (v.2.12)