A comparative study at the logical level of centralised and distributed recovery in clusters

Cluster systems are becoming more prevalent in today’s computer society and users are beginning to request that these systems be reliable. Currently, most clusters have been designed to provide high performance at the cost of providing little to no reliability. To combat this, this report looks at how a recovery facility, based on either a centralised or distributed approach could be implemented into a cluster that is supported by a check pointing facility. This recovery facility can then recover failed user processes by using checkpoints of the processes that have been taken during failure free execution.

The execution of Van Nguyen: an Australian perspective

The execution of 25-year-old Melbourne man, Van Nguyen, by Singaporean authorities on 2 December 2005 for attempting to smuggle 400 grams of heroin out of Singapore was cruel. It was also futile. Yet, there are three important lessons that can be learned from his killing. The first lesson is that if Australia is to exert genuine moral pressure on nations to abolish the death penalty it must do so in a principled manner, rather than making expedient pleas when Australians happen to be on the wrong end of the cruel practice. Secondly, sentencing practice in Australia, while not condoning capital punishment, is unjustifiably punitive and we should ameliorate the harshness of some of our sentencing laws. Finally, the death of Nguyen, while tragic, was no more tragic than the millions of other preventable deaths that occur daily throughout the world. The compassion displayed toward Nguyen should be used as a catalyst for enlarging our sympathy gland in relation to all preventable deaths. I now discuss these in further detail.

Efficient serial and parallel implementations of the cutting angle global optimisation technique

We examine efficient computer implementation of one method of deterministic global optimisation, the cutting angle method. In this method the objective function is approximated from values below the function with a piecewise linear auxiliary function. The global minimum of the objective function is approximated from the sequence of minima of this auxiliary function. Computing the minima of the auxiliary function is a combinatorial problem, and we show that it can be effectively parallelised. We discuss the improvements made to the serial implementation of the cutting angle method, and ways of distributing computations across multiple processors on parallel and cluster computers.

From parallel descriptions to parallel grammars: unravelling the complexity of Lutyen's country house folio

Based on an analysis of a corpus of forty of Sir Edwin Lutyens's country house designs (Rollo 1997) a procedure for developing Parallel Descriptions will be discussed. This procedure will attempt to facilitate an integrated approach to the analysis of complex architectural forms and will provide a framework for investigating the development of parallel grammars. This is an approach which acknowledges that design does not necessarily involve the importance of one aspect, but rather a number of coextensive issues with emphasis on the ordering and priority of these issues periodically shifting.

Toward an operating system that supports parallel processing on nondedicated clusters

Present operating systems are not built to support parallel computing on clusters - they do not provide services to manage parallelism, i.e., to manage parallel processes and cluster resources. They do not provide support for both programming paradigms, Message Passing (MP) or Distributed Shared Memory (DSM). Due to poor operating systems, users must deal with computers of a cluster rather than to see this cluster as a single powerful computer. There is a need for cluster operating systems. We claim that it is possible to develop a cluster operating system that is able to efficiently manage parallelism, support MP and DSM and offer transparency. To substantiate this claim the first version of a cluster operating system managing parallelism and offering transparency, called GENESIS, has been developed.

An MPI implementation supported by process migration and load balancing

This report describes an implementation of MPI-1 on the GENESIS cluster operating system and compares this implementation to a UNIX based MPI implementation. The changes that were made to the implementation are compared between the two, and the advantages of porting to GENESIS are detailed. This report demonstrates how GENESIS load balancing supported by process migration improves the execution performance of an MPI application. The significance of this report is in demonstrating how these services can enhance parallel programming tools to improve performance and how future parallel programming tool design could take advantage of these services.

Cluster operating system support for parallel autonomic computing

The aim of this paper is to show a general design of autonomic elements and initial implementation of a cluster operating system that moves parallel processing on clusters to the computing mainstream using the autonomic computing vision. The significance of this solution is as follows. Autonomic Computing was identified by IBM as one of computing's Grand Challenges. The human body was used to illustrate an Autonomic Computing system that possesses self-knowledge, self-configuration, self optimization, self-healing, and self-protection, knowledge of its environment and user friendliness properties. One of the areas that could benefit from the comprehensive approach created by the autonomic computing vision is parallel processing on non-dedicated clusters. Many researchers and research groups have responded positively to the challenge by initiating research around one or two of the characteristics identified by IBM as the requirements for autonomic computing. We demonstrate here that it is possible to satisfy all Autonomic Computing characteristics.

Investigation of an optimistic based concurrency control mechanism for parallel file systems

Parallel file systems offer improved performance over traditional file systems by allowing processes to access file data, stored over a number of disks, in parallel. When this file is to be accessed by many processes at the same time for reading and writing, the problem of maintaining file consistency is introduced. The general approach to maintaining file consistency is through the use locks which grant a single process exclusive access to the file. Existing studies have found parallel file systems and associated applications rarely experience conflict, resulting in unnecessary overheads from locks being acquired. This paper examines an alternate approach to maintaining file consistency, which is based on optimistic concurrency control. This approach was shown to have a much smaller overhead when compared with traditional lock-based approaches; especially in situations when there is little contention. This paper presents the design, implementation and initial testing results of an optimistic based concurrency control mechanism.

A parallel Euler approach for large-scale biological sequence assembly

Biological sequence assembly is an essential step for sequencing the genomes of organisms. Sequence assembly is very computing intensive especially for the large-scale sequence assembly. Parallel computing is an effective way to reduce the computing time and support the assembly for large amount of biological fragments. Euler sequence assembly algorithm is an innovative algorithm proposed recently. The advantage of this algorithm is that its computing complexity is polynomial and it provides a better solution to the notorious “repeat” problem. This paper introduces the parallelization of the Euler sequence assembly algorithm. All the Genome fragments generated by whole genome shotgun (WGS) will be assembled as a whole rather than dividing them into groups which may incurs errors due to the inaccurate group partition. The implemented system can be run on supercomputers, network of workstations or even network of PC computers. The experimental results have demonstrated the performance of our system.

Robust parallel filtering for mobile agent tracking

In this paper we develop a robust method of target/mobile agent tracking involving two independent estimators with separate measurement systems. The outputs of the two estimators are combined using simple trigonometry (post-estimation data fusion) and provide a robust and reliable tracking path. We demonstrate that through the use of recent advances in robust set-value state estimation, our robust parallel filter approach performs well even when the individual filters do not. Brief comparisons with common data fusion methods are conducted in order to demonstrate the advantages of our parallel (post-estimation fusion) approach

Parallel I/O scheduling in the presence of data duplication on multiprogrammed cluster computing systems

The widespread adoption of cluster computing as a high performance computing platform has seen the growth of data intensive scientific, engineering and commercial applications such as digital libraries, climate modeling, computational chemistry, computational fluid dynamics and image repositories. However, I/O subsystem performance has not been keeping pace with processor and memory performance, and is fast becoming the dominant factor in overall system performance.  Thus, parallel I/O has become a necessity in the face of performance improvements in other areas of computing systems. This paper addresses the problem of parallel I/O scheduling on cluster computing systems in the presence of data replication.  We propose two new I/O scheduling algorithms and evaluate the relative performance of the proposed policies against two existing approaches.  Simulation results show that the proposed policies perform substantially better than the baseline policies.

Robust parallel filtering design of a swarm tracking system

Swarming networks of mobile autonomous agents require inter-agent position information in order perform various tasks. The primary control input for the majority of current control strategies is inter-agent distance information. In this paper we provide a robust parallel filter based tracking scheme that allows a mobile agent to track other multiple mobile agents. The distance, angle, and relative position is given in a direct target tracking output. This allows the mobile agent to decide which information is best suited for the particular objective