Catálogo de publicaciones - libros

TeraGrid is a collaboration of partners providing a high-performance, nationally distributed capability infrastructure for computational science. The TeraGrid team has utilized multiple surveys of user requirements to develop five-year roadmaps describing new capabilities and services, organized into several new initiatives: , , and . TeraGrid is managed by the University of Chicago and includes resources at eight partner sites (Argonne National Laboratory, Indiana University, National Center for Supercomputing Applications, Oak Ridge National Laboratory, Pittsburgh Supercomputing Center, Purdue University, San Diego Supercomputer Center, and Texas Advanced Computing Center).

TeraGrid aims to assist scientists with applications that require the combination of multiple leadership class systems- including TeraGrid storage, computing, instruments, visualization, etc. – working in concert. A team of roughly 15 staff is providing hands-on assistance to application teams pursuing TeraGrid Deep projects.

TeraGrid is a set of partnerships with peer Grid projects and prototype "science gateways" that are aimed at making TeraGrid resources available to, and tailored to, entire communities of users. Science gateways are driving policy, process, and technology standards to enable web portals, desktop applications, campus clusters, and other grid infrastructure projects to seamlessly use TeraGrid resources. Initial TeraGrid science gateway projects include community portals and desktop tools supporting life sciences and biomedicine, high-energy physics, neutron science, astronomy, nanotechnology, atmospheric and climate sciences, and environmental and emergency decision-support.

TeraGrid involves the rapid evolution of the TeraGrid software and services toward interoperability with peer Grids and campus resources. Currently TeraGrid is partnering with the Open Science Grid as well as partners in Europe (e.g. UK eScience, DEISA) and Asia-Pacific (e.g. Naregi, K*Grid).

The Globus Toolkit (GT) has been developed since the late 1990s to support the development of service-oriented distributed computing applications and infrastructures. Core GT components address, within a common framework, basic issues relating to security, resource access, resource management, data movement, resource discovery, and so forth. These components enable a broader “Globus ecosystem” of tools and components that build on, or interoperate with, core GT functionality to provide a wide range of useful application-level functions. These tools have in turn been used to develop a wide range of both “Grid” infrastructures and distributed applications. I summarize here the principal characteristics of the latest release, the Web services-based GT4, which provides significant improvements over previous releases in terms of robustness, performance, usability, documentation, standards compliance, and functionality.

The China National Grid project developed and deployed a suite of grid system software called CNGrid Software. This paper presents the features and implementation of the software suite from the viewpoints of grid system deployment, grid application developers, grid resource providers, grid system administrators, and the end users.

Grid monitoring is essential for the grid management and efficiency improvement. ChinaGrid Super Vision (CGSV) is proposed for ChinaGrid to collect status information of each entity (such as resources, services, users, jobs, Network), and provide corresponding information data query and mining services. In this paper, CGSV architecture and its components are discussed. CGSV is featured by data stream integration and adaptability to cope with dynamic measurement data and multiform query requirements. Measurement data can be accessed quickly and easily through WSRF-compliant services in CGSV. Transfer and control protocols are brought forward to facilitate data stream querying and runtime producer configuration in CGSV.

Data Grids normally deal with large data-intensive problems on geographically distributed resources; yet, most current research on performance evaluation of resource scheduling in Data Grids is based on simulation techniques, which can only consider a limited range of scenarios. In this paper, we propose a formal framework via Stochastic Petri Nets to deal with this problem. Within this framework, we model and analyze the performance of resource scheduling in Data Grids, allowing for a wide variety of job and data scheduling algorithms. As a result of our research, we can investigate more scenarios with multiple input parameters. Moreover, we can evaluate the combined effectiveness of job and data scheduling algorithms, rather than study them separately.

The development of modern science requires the equipment grid to provide a scientific collaboration research platform, which can realize remote collaboration and sharing with the key instruments and equipment in wide areas. The reliability and high efficiency of a grid service chain model are key points in creation of a grid equipment system. The -calculus as powerful process algebra has a specific advantage in modeling and testing the grid service chain model. This research investigates and improves a theoretical analysis and algorithm framework for the modeling, correctness checking and analysis of the -calculus based equipment grid service chain model. It also studies on the analysis of its logistic structure and flexible modeling for the equipment grid. It would be beneficial to open up a new space in the theoretical and applied research on grid technology and formal methodology based on cross-disciplinary cooperation.

Efficient loop scheduling on parallel and distributed systems depends mostly on load balancing, especially on heterogeneous PC-based cluster and grid computing environments. In this paper, a general approach, named Performance-Based Parallel Loop Self-Scheduling (PPLSS), was given to partition workload according to performance of grid nodes. This approach was applied to three types of application programs, which were executed on a testbed grid. Experimental results showed that our approach could execute efficiently for most scheduling parameters when estimation of node performance was accurate.

With the rapid advances in Internet and Grid technique, an increasing number of applications will involve computing systems. These applications in turn create an increasing demand for efficient resource management, request handling policies and admission control. In this paper, we propose an efficient admission control algorithm to protect the critical resource of server and improve the performance of the computing system. Stability of CPU utilization is aimed to protect the server from overload and under-load. It is then beneficial to keep a satisfactory response time of requests, high throughput and less potential loss of service. We analyze the stability in detail and present a method for tuning control gains in order to guarantee the system stability. Finally, we perform simulations to evaluate the performance of the proposed algorithm. Simulation results demonstrate that the proposed algorithm stabilizes the utilization of CPU in the computing system if the control gains are appropriately chosen on the basis of system stability. It then achieves satisfactory performance.

With XML widely used in distributed system, the existing parser models, DOM and SAX, are inefficient and resource intensive for applications with large XML documents. This paper presents an index-based parser model (IBP), which contains validation and non-validation modes, supports nearly all the XML characteristics. IBP has the characters of speediness, robustness and low resource requirement, which is more suitable for mass information parsing. We presents the application of IBP in a real-time distributed monitoring prototype system, the results have shown IBP effective.

In this paper, we introduce a model for deployment and hosting of a hierarchical gird service wherein the service provider must pay to a resource provider for the use of resources. Our model produces policies that balance the number of required resources with the desire to keep the cost of hosting the service to a minimum. In each layer of our framework, we quantify the cost increase of reserved resources caused by the fluctuation of the users’ demand. A stochastic control algorithm is cast in order to resolve the problem. The results show that the model makes good decisions in the face of such uncertainties as random demand for the service.