Larry Hughes

Canadian greenhouse gas emissions: 1990–2000

Abstract The recent Berlin conference on the changing atmosphere has highlighted two major problems: first, the planets atmospheric chemistry is undergoing radical and potentially dangerous changes from the anthropomorphic emission of a variety of gases; and second, despite promises and treaty obligations, a number of countries (including the United States, Australia and Canada) are hindering efforts to stabilize these emissions at 1990 levels by the year 2000. An examination of the Canadian governments own data for carbon dioxide and methane emissions from energy sources indicates why the Canadian stabilization target cannot be met: gross emissions are increasing, per capita emissions are increasing, and emissions in terms of gross domestic product are showing minimal change.

An efficient coverage-based flooding scheme for geocasting in mobile ad hoc networks

Geocasting is a variation of traditional multicasting in which the goal is to deliver a packet to all nodes within a geographical area often referred to as a geocast region. Several protocols have been designed for geocasting in mobile ad hoc networks (MANETs) which focus on how to route packets to the geocast region efficiently, while packet delivery within the geocast region is usually achieved by flooding. Given the known limitations of simple flooding on a network, it is beneficial to look for solutions that improve the packet delivery mechanism within a geocast region. In this paper, a protocol is proposed that extends existing geocast protocols by supporting a novel packet delivery mechanism through the use of efficient flooding. The solution employs coverage information and partitioning the geocast region into grids. Using coverage information, only nodes that can reach the new portion of the region, which has not been covered by a transmission, will rebroadcast the packet. The paper shows that this solution is superior to simple flooding since it reduces the number of redundant retransmissions within the geocast region considerably.

Cartesian Ad Hoc Routing Protocols

As ad hoc networks gain in popularity, some of their limitations are becoming apparent, notably power and bandwidth restrictions. Consequently, it is necessary to utilize protocols that reduce power consumption, reduce traffic, and restrict flooding. In this paper, two adaptive, connectionless protocols and their supporting subsystems are described. The protocols, when used with directional antennas, can reduce the number of nodes involved in a transmission, thereby addressing the issue of power consumption and bandwidth utilization.

Optimizing interprocess communication for best performance in real-time systems

This paper describes a new approach in designing communications between processes in a multi-threaded environment. This technique supports real-time performance and can be adapted for embedded operating systems. At present, Message-Passing is a widely used technique in designing Interprocess Communication Systems. This design may degrade performance of Realtime Systems. The present research introduces a library-based architecture for Interprocess Communication Systems (IPC) adaptable in Real-time Embedded Operating Systems. It achieves improved real-time performance by running IPC as a set of library functions.

An Applied Approach to Teaching the Fundamentals of Operating Systems

Operating systems is an example of a subject that, with the right tools, can be taught in an applied manner, giving students the opportunity to understand, and hence appreciate, the fundamentals of commercial operating systems such as Unix and Windows. Not surprisingly, the tools needed in an operating systems course are, at a minimum, a hardware testbed and a programming language into which algorithms can be translated. This paper describes how many of the salient points concerning operating systems, such as concurrency and critical regions, can be covered in an applied manner using the PC, the C programming language, and extensions to C.

Implosion-avoidance protocols for reliable group communications

A reliable group communication protocol requires that members of the group acknowledge the reception of messages. The uncontrolled transmission of acknowledgements from the members of a group can lead to implosion; a condition that occurs when the responses flood the network or overwhelm the resources of the clients host. Scaling, the increase in the membership of a group, is a typical cause of implosion and the subsequent failure of a group communication. The problem of implosion is reviewed in this paper and a number of solutions considered; from this examination, two reliable group communication protocols are developed. Although these protocols are designed to minimize the likelihood of implosion, they also exhibit shorter data transfer times and generate less network traffic as compared to other group communication methods.<<ETX>>

Object identification in the Lego kernel

In a message‐based operating system, the kernel is responsible for forwarding messages to destination objects using a destination identifier. Part of the kernel overhead in processing the message involves mapping the identifier into a network address. This paper describes how the choice of identifier structure has simplified the techniques used by the Lego kernel when handling messages, notably: minimizing the overheads associated with address mapping, eliminating the reception of unwanted network packets, while permitting both individual (unicast) and group (multicast) communications.

The Technical Potential for Off-Peak Electricity to Serve as Backup in Wind-electric Thermal Storage Systems

Despite its potential as a secure and environmentally benign source of electricity, winds intermittency is proving to be a challenge for many electricity suppliers. One approach to overcoming this intermittency is to match it with a load that can be made to follow the wind, such as electric thermal storage systems for space heating. In such configurations, wind-generated electricity can be used for space heating and, if sufficient surplus remains, for recharging the thermal storage system. When there is a demand for heat but no wind available, the thermal storage system can discharge, meeting the space heating requirements. In extreme cases, when the thermal storage system is fully discharged and there is no wind, some form of backup energy source is required. This article examines the technical potential of off-peak electricity to ensure that wind-charged thermal storage systems are able to bridge periods of insufficient wind. The simulations show that wind heating with off-peak backup can reduce surplus electricity generated from the wind and greenhouse gas emissions. The benefits as well as the limitations of the approach are discussed.

The design and implementation of a Cartesian router

The increasing popularity of the Internet and networking has resulted in a significant growth in Internet traffic, coupled with an increase in the number of Internet routers. The increase in routers has resulted in the development of more complex routing algorithms, larger routing tables (requiring more memory), ultimately increasing the time required to search the lookup table. The Cartesian network is an attempt to overcome these problems. Instead of improving the search algorithm, it entirely removes the need for a table lookup. The Cartesian unicast routing algorithm is a novel routing methodology in which a packets route is determined by the position of the router relative to that of the destination. This paper describes the hardware design, development, and implementation of the Cartesian routers. A parallel architecture is proposed for the Cartesian routers. Field programmable gate arrays (FPGA) devices are selected as a target platform for hardware implementation.

Identifying migrated objects using mulitcast addresses

Abstract All objects in a distributed system must be associated with some form of identifier if they are to communicate with other objects. Should an object migrate between a pair of processors, it then becomes necessary to institute measures to ensure that communication can continue. Seven object identification techniques are examined here, and their application to object migration and communication is considered. The use of multicast addresses to identify objects is then proposed. Data from an experimental Ethernet illustrates the benefits of using multicast addresses, including a reduction in identifier management overheads, lower transmission times, and less network traffic.