Steve Dowers

From stand-alone programs towards grid-aware services and components: a case study in agricultural modelling with interpolated climate data

Abstract We identify three key issues to be taken into account when designing the next generation of software environments for agricultural modelling. There is a burgeoning need to support collaborative research in a search for answers to big research questions, to integrate the work of data providers and model developers and to provide more generic systems. We describe the concepts of software design of a framework, designed with these points in mind, which facilitates the integration of point-based agricultural models with methods to interpolate climate data. Our approach allows the inter-working of model and interpolation through Fortran functions that are invoked from a central framework. We advocate that the framework code remains open to collaborators, such that it may be adapted to different classes of application, whilst recognising that some module developers need to retain their control on individual elements of the software. The rationale presented within the paper continues a major move away from the stand-alone programs that still dominate agricultural models and interpolation methods. Secondly, the paper considers how these approaches are extendable to exploit opportunities in the emerging Web Service and Grid context. The emerging technology of the Grid allows geographically distributed resources in hardware, software, data and network to be co-ordinated to meet the needs of “virtual organisations.” We explore how the modularity of our existing code can be exploited in the Grid environment, whilst noting the pre-requisite of a co-operative culture in which both software developers and data providers seek to deliver services to the widest possible community of users.

Parallel processing for geographical applications: A layered approach

Abstract. Significant trends in the processing of geographical data require increasingly powerful software and hardware, consistent with the exploitation of parallel computing. Despite recent progress in technology, exploiting parallel processing is still difficult so that few applications have been developed in the environmental and geographical domains. Key issues which must be addressed in the design of parallel geographical software are described with reference to designs for three examples which use grid and raster data. The implications for parallel processing with vector-topological data are then explored. The emphasis is upon MIMD architectures using strategies of decomposition into subareas, and upon the need to facilitate development of parallel geographical applications by encapsulating the parallelism in a low-level layer of software, forming a skeletal framework upon which application algorithms can be built. The parallel layer will support distribution of datasets across the multiple processors, and the creation and collation of datasets from those processors.

Towards a framework for High Performance Geocomputation: Handling Vector Topology within a Distributed Service Environment

This paper lays out a framework, based on the emerging Open GIS standards, which will allow the integration of parallel computing technology such that it becomes a viable component of a new generation of geographical information system (GIS) software. The significant costs of parallel re-implementation have thus far acted as a major disincentive to software vendors taking advantage of parallel technology to solve performance problems. These problems will be thrown into sharp focus by the demands of web-based geographical information services. Designs for a series of software libraries, which are subject to a prototype implementation involving the use of a sophisticated data format (Neutral Transfer Format Level 4), are examined with a view to re-implementation making use of the Open GIS Abstract Specification Model. A range of services are envisaged, which can provide functions at various levels from data retrieval, spatial analysis and map generation to specialist environmental models, which are made available over the Internet. Parallelism is seen as an important route for accelerating individual transactions. These services can equally be based on large specialised parallel servers or a co-operating set of under-used workstations. The implementation strategy involves insulating standard serial algorithms from parallelism through support libraries. These libraries handle, for example, the decomposition of the data, thus effectively encapsulating the parallelism within one component of the software and allowing the creation of high-performance software components which are compatible with the Open GIS service architecture.

Towards a HPC Framework for Integrated Processing of Geographical Data: Encapsulating the Complexity of Parallel Algorithms

High-performance computing (HPC) techniques are still considered an esoteric research branch of GI processing. They are complex to use, deterring both academic modellers and commercial software developers. Yet the use of many environmental models is constrained by computation times. Furthermore, as remote sensing, environmental modelling and GIS converge, so the need for parallel computing becomes more apparent. Several case studies, parallelising the processing of raster, grid and vector-topology, demonstrate that scope exists for encapsulating the complexity of the parallelism in software frameworks, with strategies of spatial decomposition into sub-areas maximising the re-use of code from sequential algorithms. We show that parallel software frameworks can speed both the development and the execution of new applications. Based upon these case studies, the parallelisation of both interpolation and modelling in one software system is considered, with reference to pest infestation models, using both task and data Transactions in GIS, 2000, 4(3): 245±262 ß 2000 Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA. Address for correspondence: Michael J. Mineter, Parallel Architectures Laboratory for GIS, Department of Geography, University of Edinburgh, Drummond Street, Edinburgh EH8 9XP, United Kingdom. E-mail: mjm@geo.ed.ac.uk parallelism. We discuss some of the requirements of a parallel software framework to underpin the integrated analysis of geographical data and environmental models.

Partitioning of Vector-Topological Data for Parallel GIS Operations: Assessment and Performance Analysis

Geographical Information Systems (GIS) are able to manipulate spatial data. Such spatial data can be available in a variety of formats, one of the most important of which is the vector-topological. This format retains the topological relationships between geographical features and is commonly used in a range of geographical data analyses. This paper describes the implementation and performance of a parallel data partitioning algorithm for the input of vector-topological data to parallel processes.

The Performance of Vector Oriented Data Storage Strategies in ESRI's ArcGIS

The emergence of Geographical Information Systems (GIS) as an important tool in the analysis of spatial phenomena has been mirrored by the evolution of the data models underpinning such systems. When considering vector-based solutions, such developments have seen a migration from single-user, file-based, topological hybrid models to multi-user database management system (DBMS) based integrated formats, often with no inherent topology. With all these solutions still being readily available, the decision of which to employ for a given application is a complicated one. This study analyses the performance of a number of vector data storage formats for use with ESRIs ArcGIS, with the aim to facilitate the ‘intelligent selection’ of an appropriate solution. Such a solution will depend upon the application domain and both single-user and multi-user (corporate) scenarios are considered. Findings indicate that single-user ESRI coverages and multi-user ArSDE/Oracle strategies perform better when considering the range of GIS operations used to evaluate data store performance.