Catálogo de publicaciones - libros

In today’s large collaborative environments, potentially composed of multiple distinct organisations, uniform controlled access to data has become a key requirement if these organisations are to work together as . We refer to such an integrated set of as a . The (OGSA-DAI) project was established to produce a common middleware solution, aligned with the Global Grid Forum’s (GGF) OGSA vision [OGSA] to allow uniform access to data resources using a service based architecture. In this paper the service infrastructure provided by OGSA-DAI is presented providing a snapshot of its current state, in an evolutionary process, which is attempting to build infrastructure to allow easy integration and access to distributed data using grids or web services. More information about OGSA-DAI is available from the project web site: .

The data management problems in grid computing are often challenging in many aspects such as data volumes, heterogeneity, structural complexity and semantic content. Thus, e-Scientists and scientific application developers stand to benefit from tools and environments that either hide, or help to manage, the inherent complexity involved in accessing and making concerted use of the diverse resources. This paper describes OGSA-DQP, a high level data integration tool for service-based grids, and illustrates how it can be used to support grid users, via an example scientific study in bioinformatics. The paper also discusses various options for employing OGSA-DQP to handle data integration tasks as service orchestrations involving both data and analysis services.

An agent-based architecture of an active Digital Library (DL) is first described, to illustrate how electronic service provision can be supported through the use of agents. The use of mobile agents is presented as a key enabler for allowing services to be combined from a variety of providers, each of which provide a subset of the total required service. Load balancing approaches are then used to illustrate how particular performance criteria can be achieved in service provision. Extrapolation of the approach to the general Service-Oriented computing model is also discussed. A DL composed of multi-spectral imagery of the Earth, as part of the Synthetic Aperture Radar Atlas (SARA) is then used to illustrate the concepts described. The load balancing technique proposed is based on a combination of the state and model-based approaches. Experimental results demonstrating the distribution of agent load among the servers that constitute the DL, and the optimization of performance provided by the adaptability of the model employed is presented. Such an approach is particularly suited to Grid environments, which can involve a composition of services from a variety of distributed data resources.

In data intensive sciences like High Energy Physics, large amounts of data are typically distributed and/or replicated to several sites. Although there exist various ways to store and access this data within a Grid environment, site security policies often prohibit end user access to remote sites. These limitations are typically overcome using a proxy service that requires limited network connections to and from remote sites. We present a novel proxy server for the xrootd data server that provides access to data stored in an object-oriented data store (typically in ROOT format). Our proxy service operates in a structured peer-to-peer environment and allows for fault tolerant and reliable access to remote data.

One of the major advantages of Grid Computing (GC) technology is the use of geographically distributed resources. Nevertheless, just like any kind of systems, GC environments have a great problem: I/O system is the bottleneck of the whole system. In order to obtain a better performance, it is necessary and advisable to improve the data access. This problem could be solved by introducing a parallel data access to grid resources.

Since GC consists of many resources and some of them are clusters, it would be possible to exploit the parallelism among the different nodes of each cluster. We propose to use two levels of parallelism in a Grid environment to improve the data access and therefore the whole system performance. The low level will be represented by the nodes of each cluster, and the high level will include all the clusters. This paper shows a new architecture for grids, which is based on the parallel file system MAPFS, designed for high performance clusters.

Most works on Grids have taken an approach where the system is a mixture of clusters and other resources put together with the help of some services. But this solution is a simplistic one that tries to grow from the cluster perspective. We think that the Grid model should be different and near to the p2p model, especially in the I/O field where the network and the heterogeneity of the infrastructure play an important role. In this paper we present a model to organize the DataGrid Infrastructure using concepts as data phases and a p2p approach, in order to select the adequate working policies. These concepts allow the definition of a clearer model for our DataGrid Architecture than a mixture of resources. We present a model relying on the former concepts, their implementation in an I/O middleware for Grids, called GridExpand, and the evaluation of some of the concepts presented.

Bioinformatics applications manage complex biological data stored into distributed and often heterogeneous databases and require large computing power. Among these, protein structure comparison applications exhibit complex workflow structure, access different databases, require high computing power. Thus they could benefit of semantic modelling and Grid infrastructure. We present the modelling and development of the PROuST structure comparison application on the Grid using PROTEUS, a Grid-based Problem Solving Environment.

This paper describes the increasing role of ontologies in the context of Grid computing for obtaining, comparing and analyzing distributed heterogeneous scientific data. In the communities of people committed to a common goal, the management of resources and services becomes very important. We chose the application domain of human disease research and control. A characteristic of the domain is that trusted databases exist but their schemas are often poorly or not documented. The network of biomedical databases forms a loose federation of autonomous, distributed, heterogeneous data repositories ripe for information integration. Grid services will provide a dynamic way to use resources in such a large distributed scientific environment while the use of ontology enables the system to carryout reasoning at 3 levels: a) available information in all Bio-Databases (Grid nodes) worldwide, b) reasoning about the retrievable information from each node, c) reasoning about the retrieved information and presenting it in a meaningful format for users. We adopted the ontology design methodology of DOGMA and developed Generic Human Disease Ontology (GenDO) that contains common general information regarding human diseases. The information is represented in 4 “dimensions”: (a) disease types, (b) causes (c) symptoms and (d) treatments. We illustrate how this GenDO helps to produce Specific Human Disease Ontologies (SpeDO) on request. We show how the combination of two different but complementary techniques, namely Grid computing and ontology, results in a dynamic and intelligent information system. The two approaches together, being complementary, enable the system as a whole.

This paper describes the design and implementation of a model of how to integrate sensors and devices into a Grid infrastructure. We describe its proxy-based approach, the port-type requirements and the set of tools implemented to facilitate configuration of experimental scenarios. Two real world devices, a wearable medical jacket and an Antarctic lake probe, deployed out in the field using this architecture are described, along with their relevance in scientific research.

Grids constitute a promising platform to execute loosely coupled applications, which arise naturally in many scientific and engineering fields like bioinformatics, computational fluid dynamics, particle physics, etc. In this paper, we describe our experiences in porting three scientific production codes to the Grid. Those codes follow typical computational models, namely: embarrassingly distributed and master-worker. In spite of their relatively simple computational structure, consisting of many “independent” tasks, their reliable and efficient execution on computational Grids involves several issues, due to both the dynamic nature of the Grid itself and the execution and programming requirements of the applications. The applications have been developed by using the DRMAA (Distributed Resource Management Application API) interface. DRMAA routines are supported by the functionality offered by the Griday framework, that provides the runtime mechanisms needed for transparently executing jobs on a dynamic Grid environment. The experiments have been performed on Globus-based research testbeds that span heterogeneous resources in different institutions.