Peter Brezany

Novel mediator architectures for Grid information systems

Virtualization is one of the key features of the Grid. As data-intensive applications gain on importance, it becomes crucial for the success of the Grid to provide transparent access to distributed/heterogeneous data sources as well. In this paper, we describe our concepts to provide a Virtual Data Source (VDS) - a logically single homogeneous data source - on the Grid. Various data sources like relational/XML databases and comma separated value files can be combined via a Mapping Schema. By allowing to invoke user defined functions during the mediation a flexible way of resolving heterogeneities is provided. To show the feasibility of the developed concepts and elaborate the basis for a dynamic version they have been seamlessly integrated into OGSA-DAI to provide a non-proprietary, centralized and easy to integrate solution.

GridMiner: A Fundamental Infrastructure for Building Intelligent Grid Systems

The grid is considered as a crucial technology for the future knowledge-based economy and science. The Wisdom Grid project (a joint research effort of the University of Vienna and the Vienna University of Technology) aims, as the first research effort, to cover all aspects of the knowledge life cycle on the grid - from discovery in grid data repositories, to processing, sharing and finally reusing of knowledge as input for a new discovery. This paper first outlines the architecture of the Wisdom Grid infrastructure and then focuses on the kernel architecture component called Grid-Miner, which realizes the knowledge discovery, based on data mining and on-line analytical processing (OLAP) in grid repositories. A running GridMiner prototype is already available to the scientific community as an open service system.

On-line analytical processing on large databases managed by computational grids

Management of large data repositories integrated into the grid poses new challenges for grid research. There already exist several successful data grid projects addressing processing files storing large volumes of scientific data and projects developing services for accessing remote relational and XML databases. However, so far, no effort was devoted to on-line analytical processing (OLAP), an essential support for modern decision support systems. We present our approach to the design and implementation of a grid-enabled OLAP server, which is one functional building block of the GridMiner system, a novel infrastructure for knowledge discovery in grid databases. We present the global architecture model of our solution, describe how the OLAP components are integrated into the GridMiner system, and present the software architecture of our first prototype. The OLAP components were implemented in Java on top of the software toolkit Globus 3.

Concurrent file operations in a high performance FORTRAN

The authors propose constructs to specify I/O (input/output) operations for distributed data structures in the context of Vienna FORTRAN. These operations can be used by the programmer to provide information which will allow the compiler and runtime environment to optimize the transfer of data to and from secondary storage. Although the language constructs presented have been proposed in the context of Vienna FORTRAN, they can be easily integrated into any other high-performance FORTRAN extension.<<ETX>>

Workflow enactment engine for WSRF-compliant services orchestration

The convergence of grid services more closely to the Web services in the recent years has opened new challenges in designing modern system architectures. Scientific applications based on the grid service architectures can be now viewed as complex processes utilizing available resources so that they co-operate to implement the desired behavior of the system. The workflows became a natural method for describing such processes typically executed on various dispersed resources. Moreover, a number of new specifications and standards is continuously being developed to support the unavoidable and highly demanded interoperability between the services composing the processes. In order to orchestrate the grid services, a workflow engine able to cope with the different standards is highly desirable. The work presented in this paper is carried out in the context of the Globus Incubator project named Workflow Enactment Engine Project. The project was established to provide an easy-to-use and manage workflow enactment engine for grid services orchestration using the WS-BPEL 2.0 language.

Toward a Grid-Based Zero-Latency Data Warehousing Implementation for Continuous Data Streams Processing

Continuous data streams are information sources in which data arrives in high volume in unpredictable rapid bursts. Processing data streams is a challenging task due to (1) the problem of random access to fast and large data streams using present storage technologies and (2) the exact answers from data streams often being too expensive. A framework of building a Grid-based Zero-Latency Data Stream Warehouse (GZLDSWH) to overcome the resource limitation issues in data stream processing without using approximation approaches is specified. The GZLDSWH is built upon a set of Open Grid Service Infrastructure (OGSI)-based services and Globus Toolkit 3 (GT3) with the capability of capturing and storing continuous data streams, performing analytical processing, and reacting autonomously in near real time to some kinds of events based on a well-established knowledge base. The requirements of a GZLDSWH, its Grid-based conceptual architecture, and the operations of its service are described in this paper. Furthermore, several challenges and issues in building a GZLDSWH, such as the Dynamic Collaboration Model between the Grid services, the Analytical Model, and the Design and Evaluation aspects of the Knowledge Base Rules are discussed and investigated.

SUPERB support for irregular scientific computations

Runtime support for parallelization of scientific programs is needed when some information important for decisions in this process cannot be accurately derived at compile time. This paper describes a project which integrates runtime parallelization with the advanced compile-time parallelization techniques of SUPERB. Besides the description of implementation techniques, language constructs are proposed, providing means for the specification of irregular computations. SUPERB is an interactive SIMD/MIMD parallelizing system for the Suprenum, iPSC/860 and Genesis-P machines. The implementation of the runtime parallelization is based on the Parti procedures developed at ICASE NASA.<<ETX>>

A Software Architecture for Massively Parallel Input-Output

For an increasing number of data intensive scientific applications, parallel I/O concepts are a major performance issue. Tackling this issue, we provide an outline of an input/output system designed for highly efficient, scalable and conveniently usable parallel I/O on distributed memory systems. The main focus of this paper is the parallel I/O runtime system support provided for software-generated programs produced by parallelizing compilers in the context of High Performance FORTRAN efforts. Specifically, our design is presented in the context of the Vienna Fortran Compilation System.

Towards Realization of Dataspaces

Database management systems providing powerful data management services have been developed in the recent decades to a great extent. All these systems are based on its own single data model, whereas the demand for managing multiple data sources with different data models is rapidly expanding. Therefore the need for intelligent management systems providing access to those heterogeneous and often distributed data sources and allowing to search and query them as a single information source, has never been greater. This research challenge is faced by the vision of dataspaces. So far, no effort was devoted to realization of dataspace concepts on the grid. In this paper, we propose the architecture of a dataspace management system and discuss how some current components of the grid technology can support the future implementations of such architecture

Knowledge grid support for treatment of traumatic brain injury victims

Traumatic brain injuries (TBIs) typically result from accidents in which the head strikes an object. Among all traumatic causes of death, the TBI is the most serious one. Moreover, TBI can also significantly affect many cognitive, physical, and psychological skills of the cases that survive. It is clear from the data presented in many international studies, that improvement of outcomes is useful not only for the patients, but also for entire national economy. This paper introduces a new software infrastructure, called TBI Knowledge Grid that will provide health professionals involved in the treatment and management of TBI patients with Information Technology tools, which will enable them to discover and access relevant knowledge and information from different distributed and heterogeneous data sources.