Creating scientific workflow applications is a very challenging task due to the complexity of the distributed computing environments involved, the complex control and data flow requirements of scientific applications, and the lack of high-level languages and tools support. Particularly, sophisticated expertise in distributed computing is commonly required to determine the software entities to perform computations of workflow tasks, the computers on which workflow tasks are to be executed, the actual execution order of workflow tasks, and the data transfer between them. Qin and Fahringer present a novel workflow language called Abstract Workflow Description Language (AWDL) and the corresponding standards-based, knowledge-enabled tool support, which simplifies the development of scientific workflow applications. AWDL is an XML-based language for describing scientific workflow applications at a high level of abstraction. It is designed in a way that allows users to concentrate on specifying such workflow applications without dealing with either the complexity of distributed computing environments or any specific implementation technology. This research monograph is organized into five parts: overview, programming, optimization, synthesis, and conclusion, and is complemented by an appendix and an extensive reference list. The topics covered in this book will be of interest to both computer science researchers (e.g. in distributed programming, grid computing, or large-scale scientific applications) and domain scientists who need to apply workflow technologies in their work, as well as engineers who want to develop distributed and high-throughput workflow applications, languages and tools.

Automatic Performance Prediction of Parallel Programs presents a unified approach to the problem of automatically estimating the performance of parallel computer programs. The author focuses primarily on distributed memory multiprocessor systems, although large portions of the analysis can be applied to shared memory architectures as well. The author introduces a novel and very practical approach for predicting some of the most important performance parameters of parallel programs, including work distribution, number of transfers, amount of data transferred, network contention, transfer time, computation time and number of cache misses. This approach is based on advanced compiler analysis that carefully examines loop iteration spaces, procedure calls, array subscript expressions, communication patterns, data distributions and optimizing code transformations at the program level; and the most important machine specific parameters including cache characteristics, communication network indices, and benchmark data for computational operations at the machine level. The material has been fully implemented as part of P3T, which is an integrated automatic performance estimator of the Vienna Fortran Compilation System (VFCS), a state-of-the-art parallelizing compiler for Fortran77, Vienna Fortran and a subset of High Performance Fortran (HPF) programs. A large number of experiments using realistic HPF and Vienna Fortran code examples demonstrate highly accurate performance estimates, and the ability of the described performance prediction approach to successfully guide both programmer and compiler in parallelizing and optimizing parallel programs. A graphical user interface is described and displayed that visualizes each program source line together with the corresponding parameter values. P3T uses color-coded performance visualization to immediately identify hot spots in the parallel program. Performance data can be filtered and displayed at various levels of detail. Colors displayed by the graphical user interface are visualized in greyscale. Automatic Performance Prediction of Parallel Programs also includes coverage of fundamental problems of automatic parallelization for distributed memory multicomputers, a description of the basic parallelization strategy and a large variety of optimizing code transformations as included under VFCS.

Distributed and Parallel Systems: From Cluster to Grid Computing, is an edited volume based on DAPSYS 2006, the 6th Austrian-Hungarian Workshop on Distributed and Parallel Systems, which is dedicated to all aspects of distributed and parallel computing. The workshop was held in conjunction with the 2nd Austrian Grid Symposium in Innsbruck, Austria in September 2006. This book is designed for a professional audience composed of practitioners and researchers in industry. It is also suitable for advanced-level students in computer science.

In a dynamic computing environment, such as the Grid, resource management plays a crucial role for making distributed resources available on-demand to anyone from anywhere at any time without undermining the resource autonomy; this becomes an art when dealing with heterogeneous resources distributed under multiple trust domains spanning across the Internet. Today Grid execution environments provide abstract workflow descriptions that need a dynamic mapping to actual deployments; this further accentuates the importance of resource management in the Grid.

This monograph renders boundaries of the Grid resource management, identifies research challenges and proposes new solutions with innovative techniques for on-demand provisioning, automatic deployments, dynamic synthesis, negotiation-based advance reservation and capacity planning of Grid resources. The Grid capacity planning is performed with multi-constrained optimized resource allocations by modelling resource allocation as an on-line strip packing problem and introducing a new solution that optimizes resource utilization and QoS while generating contention-free solutions.

On-demand resource provisioning becomes possible by simplifying abstract resource descriptions independent from the concrete installations. The book further explains the use of the semantic web technologies in the Grid to specify explicit definitions and unambiguous machine interpretable resource descriptions for intelligent resource matching and synthesis; the synthesis process generates new compound resources with aggregated capabilities and prowess. The newly introduced techniques haven been developed and integrated in ASKALON Grid application development and runtime environment, deployed in the Austrian Grid, and demonstrated through well performed experiments.

The Grid computing concept, which allows users to integrate administratively and g- graphically dispersed computing resources, has been gaining traction in a number of application areas during the past few years. By interconnecting many - heterogeneous, though usually virtualized - computing resources, virtual computer centers or superc- puters can be created, providing a seamless supply of computing resources. Grid comp- ing provides benefits not only for scientific computing (e.g., SETI@home, which interconnects one million computers across 226 countries with a total processing power of 711 TFLOPS) but also in a commercial environment. It is projected that computing Grids can lower the total IT costs of businesses by 30%. The report "Grid Computing: A Vertical Market Perspective 2005-2010" (by The Insight Research Corporation) estimates an increase of worldwide Grid spending from $714.9 million in 2005 to approximately $19.2 billion in 2010. One of the most prominent activities in academia is the EGEE project being funded with 30 MEuro by the European Commission. EGEE brings together researchers from over 27 countries with the common aim of developing a service Grid infrastructure, which is suited for scientific computing with very high demand for processing power.

From Cluster to Grid Computing is an edited volume based on DAPSYS 2006, the 6th Austrian-Hungarian Workshop on Distributed and Parallel Systems, which is dedicated to all aspects of distributed and parallel computing. The workshop was held in conjunction with the 2nd Austrian Grid Symposium in Innsbruck, Austria in September 2006. Distributed and Parallel Systems: From Cluster to Grid Computing is designed for a professional audience composed of practitioners and researchers in industry. This book is also suitable for advanced-level students in computer science.

Grid computing has become a topic of significant interest in the scientific community as a means of enabling application developers to aggregate resources scattered around the globe for solving large-scale scientific problems. This monograph addresses four critical software development aspects for the engineering and execution of applications on parallel and Grid architectures.

A new directive-based language called ZEN is proposed for compact specification of wide value ranges of interest for arbitrary application parameters, including problem or machine sizes, array or loop distributions, software libraries, interconnection networks, or target execution machines. Based on the ZEN language, a novel experiment management tool called ZENTURIO is developed for automatic experiment management of large-scale performance and parameter studies on parallel and Grid architectures. This tool has been validated with respect to functionality and usefulness on several real-world parallel applications from various domains, including theoretical chemistry, photonics, finances, and numerical mathematics.

Depending on the ZENTURIO experiment management architecture a generic optimization framework is built up that integrates general-purpose meta-heuristics for solving NP-complete performance and parameter optimization problems in an exponential search space specified using the ZEN experiment specification language. Finally a timely approach is proposed for modeling and executing scientific workflows in dynamic and heterogeneous Grid environments, introducing an abstract formal model for hierarchical representation of complex directed graph-based workflows.

Thus this monograph contributes to various research areas related to integrated tool development for efficient engineering and high performance execution of scientific applications in Grid environments.

Automatic Performance Prediction of Parallel Programs presents a unified approach to the problem of automatically estimating the performance of parallel computer programs. The author focuses primarily on distributed memory multiprocessor systems, although large portions of the analysis can be applied to shared memory architectures as well. The author introduces a novel and very practical approach for predicting some of the most important performance parameters of parallel programs, including work distribution, number of transfers, amount of data transferred, network contention, transfer time, computation time and number of cache misses. This approach is based on advanced compiler analysis that carefully examines loop iteration spaces, procedure calls, array subscript expressions, communication patterns, data distributions and optimizing code transformations at the program level; and the most important machine specific parameters including cache characteristics, communication network indices, and benchmark data for computational operations at the machine level. The material has been fully implemented as part of P3T, which is an integrated automatic performance estimator of the Vienna Fortran Compilation System (VFCS), a state-of-the-art parallelizing compiler for Fortran77, Vienna Fortran and a subset of High Performance Fortran (HPF) programs. A large number of experiments using realistic HPF and Vienna Fortran code examples demonstrate highly accurate performance estimates, and the ability of the described performance prediction approach to successfully guide both programmer and compiler in parallelizing and optimizing parallel programs. A graphical user interface is described and displayed that visualizes each program source line together with the corresponding parameter values. P3T uses color-coded performance visualization to immediately identify hot spots in the parallel program. Performance data can be filtered and displayed at various levels of detail. Colors displayed by the graphical user interface are visualized in greyscale. Automatic Performance Prediction of Parallel Programs also includes coverage of fundamental problems of automatic parallelization for distributed memory multicomputers, a description of the basic parallelization strategy and a large variety of optimizing code transformations as included under VFCS.