Robert D. Kent

Genetic algorithm based scheduler for computational grids

In the context of highly scalable distributed resource management architectures for grid computing, we present a genetic algorithm based scheduler. A scheduler must use the available resources efficiently, while satisfying competing and mutually conflicting goals. The grid workload may consist of multiple jobs, with quality-of-service constraints. A directed acyclic graph (DAG) represents each job, taking into account arbitrary precedence constraints and arbitrary processing time. The scheduler has been designed to be compatible with other tools being developed by our grid research group. We present the design, implementation and test results for such a scheduler in which we minimize make-span, idle time of the available computational resources, turn-around time and the specified deadlines provided by users. The architecture is hierarchical and the scheduler is usable at either the lowest or the higher tiers. It can also be used in both the intra-grid of a large organization and in a research grid consisting of large clusters, connected through a high bandwidth dedicated network.

An adaptive generalized scheduler for grid applications

The compute resources of a grid resource-service provider may be distributed over a wide geographical area. If the resource-service provider is to use his resources effectively, in addition to the characteristics of the compute-nodes and the applications, the characteristics of the communication network must also be known. A generalized scheduler should be able to handle a diverse set of jobs, with arbitrary inter-dependencies among processes and arbitrary communication channel delays. On a grid, the scheduling algorithm should respond quickly to the changing workload and environmental conditions without causing much overhead. Hence it should be able to customize its strategy in accordance with the prevailing conditions. In this paper, the algorithm for an adaptive scheduler, which can be used to map a set of jobs, of wide diversity, to a dynamic set of nodes, with prior reservations, is being presented. The scheduler has been tested extensively.

PKI-Based Authentication Mechanisms in Grid Systems

Grids have emerged as the basic infrastructure for high performance distributed computing and data collaborations. Although they depict an attractive new world of computing, security is the biggest barrier against wide adoption of Grids. Authentication is the basis of security in grids. GSI uses X.509 PKI and proxy certificates as authentication foundation, and uses gateway for mapping certificates between different authentication mechanisms. In this article, we review PKI and PKI-based authentication mechanisms used in grid systems. These mechanisms are insufficient or problematic under some circumstances. We study and analyze some prominent challenges or problems: compatibility across different PKIs, proxy certificate revocation, security weakness, and authentication in ad hoc grids. For each of them, we introduce possible solutions, and analyze state-of-the-art technologies and ongoing researches that indicate the direction of future work on this topic.

A key management and secure routing integrated framework for Mobile Ad-hoc Networks

Key management (KM) and secure routing (SR) are two most important issues for Mobile Ad-hoc Networks (MANETs), but previous solutions tend to consider them separately. This leads to KM-SR interdependency cycle problem. In this paper, we propose a KM-SR integrated scheme that addresses KM-SR interdependency cycle problem. By using identity based cryptography (IBC), this scheme provides security features including confidentiality, integrity, authentication, freshness, and non-repudiation. Compared to symmetric cryptography, traditional asymmetric cryptography and previous IBC schemes, this scheme has improvements in many aspects. We provide theoretical proof of the security of the scheme and demonstrate the efficiency of the scheme with practical simulation.

Towards a Unified Data Management and Decision Support System for Health Care

We report on progress in development of a unified data management and decision support system, UDMDSS, for application to injury prevention in health care. Our system is based on a modular architecture which supports real-time web-base desktop and mobile data acquisition, semantic data models and queries, Bayesian statistical analysis, artificial intelligence agent-based techniques to assist in modelling and simulation, subjective logic for conditional reasoning with uncertainty, advanced reporting capabilities and other features. This research work is being conducted within a multi-disciplinary team of researchers and practitioners and has been applied to a Canadian national study on child safety in automobiles and also in the context of patient falls in a hospital.

A Framework for Revocation of Proxy Certificates in a Grid

Proxy certificates (PCs) are essential in grid security infrastructure (GSI), but their revocation still remains problematic in grid environments. The mechanisms for normal end entity certificates (EEC) revocation need to be reconsidered or revised for PC revocation in grids. In this paper, we present a new framework that addresses the PC revocation problem based on MyProxy - the online credential repository - in Globus toolkit. The framework has been implemented to prove its feasibility. The implementation is light-weight and has been integrated into the Globus architecture. We consider this as an important contribution to Globus and MyProxy and of significance to users and providers of critical grid services.

Towards a “Just-in-Time” Distributed Decision Support System in Health Care Research

The application of semi-automated decision support systems in health care faces challenging tasks mainly in generating evidence-based recommendations in a short critical time window. Traditional data collection and survey methodology to generate evidence for the decision support systems also suffers from a slow turn-around time. This chapter reports on a multi-disciplinary project between computer and health sciences to introduce a cumulative framework encapsulating innovative distributed data collection methodology, coupled with an intelligent multi-agent, socially driven decision support system. We report on the current design and implementation aspects of this integrated system with a case study in injury prevention to verify the initial model.

Topology reconfiguration mechanism for traffic engineering in WDM optical network

Optical wavelength division multiplexed (WDM) networks provide an excellent transmission medium for voice and data traffic, streaming media, and high performance and grid computing needs. Rearrangeability is one compelling characteristic of WDM optical networks that allows network operators to rearrange the networks in response to changing traffic demands and element failures to provide improved network performance. Under changing traffic flows and abhorrent network conditions reconfiguration is an ongoing process. Available approaches are not able to handle extreme load conditions, traffic bursts, and element failures and, so far, these approaches cover only limited aspects of the problem of automated reconfiguration. In this work we present an adaptive reconfiguration mechanism for WDM optical networks (ARWON). Two heuristically based algorithms, combination algorithm (CA) and multi lightpath change (MLPC) algorithm, are also proposed to support implementation of ARWON. Simulation experiments covered all ranges of traffic flows and element failure scenario and performance of ARWON was validated with an established contemporary reconfiguration algorithm. Results show that under all loading conditions and link failure scenario ARWON performed better than the comparison algorithm. Under extreme loading conditions ARWON incurred 11 times less traffic loss than single lightpath change (SLPC) algorithm and on average SLPC was carrying 20 more lightpaths than ARWON, per observation cycle, that were 100 percent loaded. Under link failure scenario ARWON recovered faster and rerouted the traffic in one step, whereas SLPC recovered slowly and recovery took much longer duration than ARWON. For other loading conditions, ranging from low to high loading, it was observed that during early stages of simulation ARWON performs at par or marginally better than SLPC. Performance of ARWON steadily improves as simulation progresses over longer duration.

An integrated key management and secure routing framework for Mobile Ad-Hoc Networks

Key management (KM) and secure routing (SR) are two most important issues for Mobile Ad-hoc Networks (MANETs), but previous solutions tend to consider them separately. This leads to KM-SR interdependency cycle problem. In this paper, we propose an integrated KM-SR scheme that addresses KM-SR interdependency cycle problem. By using identity based cryptography (IBC), this scheme provides security features including confidentiality, integrity, authentication, freshness, and non-repudiation. Compared to symmetric cryptography, traditional asymmetric cryptography and previous IBC schemes, this scheme has improvements in many aspects. We provide theoretical proof of the security of the scheme and demonstrate the efficiency of the scheme with practical simulation.

A reputation model framework for artificial societies: a case study in child vehicle safety simulation

Formalizing reputation into a complex social model poses significant challenges, mainly due to its distinct social nature. In this paper we introduce the notion of reputation into the child vehicle safety simulation. From a health and safety perspective, the aim of the model is to reduce injury in children by minimizing incorrect usage of child vehicle constraints by influencing driver behaviour. A cultural framework was previously established to enable external injection of knowledge, or intervention, into the artificial society. A dynamic social network allowed the acquisition, and subsequent exchange and evolution of knowledge. We hypothesize that selective intervention criteria would achieve better system convergence. We consequently introduce reputation to be a viable selection criterion. We establish a generic reputation framework that would allow us to test alternate formalizations of reputation models. We report on the generic framework design and three initial reputation models with their respective comparative performance and potential to improve the intervention outcome.