Chris Bates

Modelling test data for performance evaluation of large parallel database machines

Parallel servers offer improved processing power for relational database systems and provide system scalability. In order to support the users of these systems, new ways of assessing the performance of such machines are required. If these assessments are to show how the machines perform under commercial workloads they need to be based upon models which have a real commercial basis. This paper shows how a realistic model of a financial application has been developed and how a set of tools has been created which allow the implementation of the model on any commercial database system. The tools allow the generation of large quantities of test data in a manner which renders it amenable to subsequent independent analysis. The test data thus generated forms the basis for the performance tuning of parallel database machines.

Scrum down: a software engineer and a sociologist explore the implementation of an agile method

This paper provides an overview and position statement on the work undertaken as part of a project to explore the implementation of Scrum in the context of an interactive digital media software development company. The project is being undertaken in the Communication and Computing Research Centre at Sheffield Hallam University.

Evaluation of High Performance Parallel Database Machines

Many organisations are considering the use of large relational databases to implement their information needs. For large organisations this will result in the implementation of a machine dedicated to that task. In the near future, these database machines will use high performance parallel engines to provide the required computing resource. Performance evaluation is particularly important when the design is being considered for future scalability and the ability of the design to cope with unexpected data access. The paper reports on the development of a synthetic or “mimic” database which models a real application: this mimic not only forms the basis for a new benchmark for large SQL databases, but can also be used to support the design process.

Benchmarking Parallel SQL Database Machines

New parallel database systems are being developed to support the data processing needs of large commercial organisations: these offer increased performance in terms of processing throughput and enhanced support for both on-line transaction processing and management information provision. In order to evaluate the performance of these systems a new SQL benchmark has been developed which models the requirements for real database applications. Based on a large commercial database the benchmark system includes the specification of the test or “mimic” database and sets of transactions and queries, and the data/transaction generation tools. The benchmark uses synthetic data which closely resembles real world data and provides for evaluation of the scalability of system under test.

Quando: Enabling Museum and Art Gallery Practitioners to Develop Interactive Digital Exhibits

Museums and Art Galleries are challenged to inspire, engage and involve visitors by presenting their collections within physical exhibitions. Curators and exhibition professionals are increasingly telling stories using digital interactivity. This work introduces Quando, a visual programming based toolset that domain experts can use to create interactive exhibits. A small case study demonstrates the language in use at during an archaeological excavation.

Students’ understanding of computational thinking with a focus on decomposition in building network simulations

This paper reports a study into students’ understanding of decomposition when building network simulations. Students were asked to complete three problem-solving tasks involving designing and troubleshooting computer networks using simulation software. Through online surveys, interviews and focus groups the students’ understanding of computational thinking was interrogated. The results show that students were not conscious that they were applying computational thinking concepts when designing and troubleshooting networks on simulation software. It appears their interest were to simply get problems solved but not necessarily with the understanding of the application of the concepts of computational thinking.

Recovering Design Patterns from Large Codebases

Understanding and using Design Patterns improves software quality through better comprehension of programs for both experienced developers and students. Often design patterns are learned through simplified “toy” programs and exemplars that foreground the structure of the pattern. However in production code the objects and methods that comprise the pattern can be hidden within a complex mesh of intra-code relationships. This paper introduces APRT, an ANTLR-based tool that recovers the structure of both static and dynamic patterns from large codebases so that they can be studied in context.

The Odyssey Project – Understanding and Implementing User Needs in the Context of Ballistic Crime Data Exchange

Gun crime – that is potential and actual illegal events in which firearms, ammunition and other ballistic items are involved – happens widely across the EU, though the levels and definitions of such events vary. Such variations come from the multiple methods, systems and legal frameworks in EU member States. The majority of law enforcement agencies (LEAs) believe that criminals actively use and move firearms across the EU. As a result criminals, guns and evidence travel across borders. Unfortunately comparable levels of data on gun crimes held by LEAs does not travel across borders, despite gun crime information being collected widely across the EU. Especially as not all the data is collected electronically and database formats, structures and data fields vary. In addition different ballistics imaging systems are used across the EU. The Odyssey Project was part-funded under the EU FP7 programme to explore the challenges of establishing a Pan-European ballistics and crime information intelligence network and to propose solutions, including a demonstrator prototype system. This chapter reviews the overall findings from the Odyssey Project and the key features of the prototype.

CREATING REALISTIC DATA FOR PERFORMANCE EVALUATION

When buying, analysing or tuning a large relational database system knowledge about performance is vital. The system tester has many different benchmarks available. These tightly specify transactions, queries and data structures but few pay any attention to the creation of data sets. This means that either data volumes are small or the data itself is lacks many of the attributes of production data. The Data Generation language (DGL) is an interpreted language which can be used to specify the structure and format of text-based relational data sets. DGL lets systems testers create realistic volumes of relational data which adhere to any database structure. DGL data even includes relational links between arbitrary tuples. This paper introduces DGL and shows how the language can be used to create data for the industry standard TPC-H decision support benchmark.

Creating dynamic content for the Web using Perl

Web development is one of the growth areas of programming. The creators of E-commerce and small hobby sites use many of the same tools and techniques. Development languages become fashionable, wildly popular then fade away to be replaced by the next big thing. Some languages have such power and suitability that they deserve to be taken seriously and used widely rather than discarded when an alternative comes along. In this paper CHRIS BATES presents the programming language Perl which is widely used on Web sites that create and present dynamic content. He focuses on the features of Perl that make it so well suited for Web development and worth considering for both experienced programmers and beginners alike. Copyright