Search

During the past few years, Grid Computing has matured in terms of programming models and available tools. Some tools like the Globus Toolkit version 2 (GT2) are used in many international high performance distributed computing projects. Recently, the OGSA standard (Open Grid Services Architecture) has been defined, proposing a radically new philosophy compared to that of GT2. Analyzing the evolution of the scientific community working on Grid Computing, we foresee a progressive shift of current developments to this new standard, that already has a reference implementation: GT3. This paper describes the analysis and design of an architecture of OGSA Grid Services that aims to integrate both platforms seamlessly, allowing remote job invocation from GT3 to GT2 holding all the security properties, and transparent for the user. [ABSTRACT FROM AUTHOR]

This paper presents design concepts and implementation overview of an Internet-sharing system, iShare. iShare supports end users as well as providers of Internet resources in disseminating, accessing and using these resources, in a way that allows open participation. A fully decentralized organization allows providers to simply post their resources on any web page, imposing no restrictions on resources attributes, administrative rules, and access protocols. Underneath its user surface it employs peer-to-peer information dissemination, advanced resource matching, open migration, and automatic service portal mechanisms. In addition to the qualitative comparison with related work, we evaluate the system in terms its efficiency of resource discovery and job execution. [ABSTRACT FROM AUTHOR]

The mobile ad hoc networks are particularly vulnerable to intrusion, as its features of open medium, dynamic changing topology, cooperative routing algorithms. The traditional way of protecting networks with firewalls and encryption software is no longer sufficient and effective for those features, because no matter how secure the mobile ad hoc networks, its is still possible the nodes are compromised and become malicious. In this paper, we propose a novel intrusion detection approach for mobile ad hoc networks by using finite state machine. We construct the finite state machine (FSM) by the way of manually abstracting the correct behaviours of the node according to the routing protocol of Dynamic Source Routing (DSR). The monitor nodes cooperatively monitor every node's behaviour by the FSM. Our approach can detect real-time attacks without signatures of intrusion or trained data. Finally, we evaluate the intrusion detection method through simulation experiments. [ABSTRACT FROM AUTHOR]

The computational grid, as it is today, is vulnerable to node failures and the probability of a node failure rapidly grows as the size of the grid increases. There have been several attempts to provide fault tolerance using checkpointing and message logging in conjunction with the MPI library. However, the Grid itself should be active in dealing with the failures. We propose a dynamic reconfigurable architecture where the applications can regroup in the face of a failure. The proposed architecture removes the single point of failure from the computational grids and provides flexibility in terms of grid configuration. [ABSTRACT FROM AUTHOR]

World Wide Web is one of the most important applications on the Internet. Net surfers can browse the Web pages by tracing the links embedded in the hypertexts. However, one will be stuck when there are no hyperlinks or the hyperlinks are broken. This paper proposes a new paradigm for Web system called WebGrid, which consists of a dynamic collection of Web servers that work cooperatively. WebGrid synchronizes the Web contents that are usually stored in the underlying database systems in order to provide a uniform view of all resources. Net surfers can retrieve all resources in the WebGrid through one local Web server. The paper addresses WebGrid architecture and some related concepts. Key issues such as the WebGrid topology, the resource synchronization mechanism and strategy, as well as the Web site customization are also discussed in the paper. [ABSTRACT FROM AUTHOR]

While existing work concentrates on developing QoS models of business workflows and Web services, few tools have been developed to support the monitoring and performance analysis of scientific workflows in Grids. This paper describes a Grid service for performance monitoring and analysis of Grid scientific workflows. The service utilizes workflow graphs and various types of performance data including monitoring data of resources, execution status of activities, and performance measurement obtained from the dynamic instrumentation, to provide a rich set of monitoring and performance analysis features. We store workflows and their relevant information, devise techniques to compare constructs of different workflows, and support multi-workflow analysis. [ABSTRACT FROM AUTHOR]

Grid computing is promissing as an infrastructure that allows users to use distributed computer resources by simple connecting a computer to the network without special operation; just like connecting to the electricity, water or gas grid. In this paper, regarding resources as services, we propose a new architecture for realizing an environment in which users can use services that are in various locations through their portable terminals. In this architecture, a service is managed by an agent, which has two resource management spaces named the Public Zone and the Private Zone. The Public Zone is a space for realizing flexible public service use. The Private Zone is a space for protecting private user information. Moreover, agents are organized in a group called the community and are managed independently in each community. Thus, we realize both of flexible service use while protecting private information. [ABSTRACT FROM AUTHOR]

A central idea of Grid Computing is the virtualization of heterogeneous resources. To meet this challenge the Institute for Scientific Computing (IWR), has started the project CampusGrid. Its medium term goal is to provide a seamless IT environment supporting the on-site research activities in Physics, Bioinformatics, Nanotechnology and Meteorology. The environment will include all kinds of HPC resources: vector computers, shared memory SMP servers and clusters of commodity components, InfiniBand-Clusters as well as a shared high-performance SAN storage solution and a global file system. The paper shows the ideas, the test-bed and informs about the current project status and scheduled development tasks. This is associated with reports on other activities in the fields of Grid computing and high performance computing at IWR and D-Grid. [ABSTRACT FROM AUTHOR]

The mapping of workflows on Grid resources differs from mapping a single job, as dependencies between the subjobs has to be considered and resolved. This is a major task, if underlying Service Level Agreements (SLAs) have to be satisfied, which define the deadline for the entire workflow but also allow flexibility while assigning the subjobs to resources. Therefore, new requirements regarding selection of optimal resource destination and satisfying multiple time constraints are important issues, which are not completely addressed by existing methods. This paper presents a method which performs an efficient and precise assignment of workflow subjobs to Grid resources with respect to SLAs and subjob dependencies. The quality of the created results is evaluated with a number of experiments and compared to the results of existing planning tools. [ABSTRACT FROM AUTHOR]

The mapping problem has been studied extensively. However, algorithms which were designed to map a parallel application on a computational grid, such as MiniMax, FastMap and genetic algorithms have shortcomings. In this paper, a new algorithm, Quick-quality Map (QM), is presented. Experimental results show that QM performs better than the other algorithms. For instance, QM can map a 10000-task parallel application on a testbed of 2992 nodes in 6.35 seconds, and gives the lowest execution time whereas MiniMax and a genetic algorithm, respectively, take approximately 1700 and 660 seconds, but produce 1.34 and 6.60 times greater execution times than QM's. [ABSTRACT FROM AUTHOR]

Detecting illegal resource access in the setting of grid computing is similar to the problem of virus detection as put forward by Fred Cohen in 1984. We discuss Cohen's impossibility result on virus detection, and introduce "risk assessment of security hazards", a notion that is decidable for a large class of program behaviors. Keywords: Malcode, Program algebra, Thread algebra, Virus, Worm. [ABSTRACT FROM AUTHOR]

This paper presents an architecture for specifying, monitoring and validating Service Level Agreements (SLA) for use in Grid environments. SLAs are an essential component in building Grid systems where commitments and assurances are specified, implemented and monitored. Targeting compute resources, an SLA manager reserves resources for user applications requiring resources on demand. Methods for automated monitoring and violation capture are discussed showing how Service Level Objectives (SLO) can be validated. A SLA for a compute service is specified and experiments carried out on the White Rose Grid. Results are presented in the form of a SLA document and show the violations that were captured during task execution. [ABSTRACT FROM AUTHOR]

The need for distributed file systems has been growing for decades to provide clients with efficient and scalable high-performance accesses to stored data. The clients physically share storage devices connected via a network like GigaEthernet or Fibre Channel and, on those clients, distributed file systems take responsibility for providing coordinated accesses and consistent views of shared data. In such a distributed computing environment, one of the major issues affecting in achieving substantial I/O performance and scalability is to build an efficient locking protocol. In this paper, we present a distributed locking protocol that enables multiple nodes to simultaneously write their data to distinct data portions of a file, while providing the consistent view of client cached data, and conclude with an evaluation of the performance of our locking protocol. Keywords: locking protocol, distributed file system, distributed computing, MPI-IO. [ABSTRACT FROM AUTHOR]

Scheduling distributed applications effectively and efficiently on Grid environments is difficult because of the dynamic and heterogeneous characteristics of the Internet. In this paper, we propose a loosely coupled application model for building distributed applications on Grids. We assume that a Grid application is composed of a group of independent modules. Each module performs either a remote service request or local processing. Different modules in such an application exchange information by explicitly described data that can be understood by both the application and the Grid environment. Each module is triggered by its input data, and finally it produces some output data. All information exchanges are completed transparently as they are carried out by the Grid management system. We call a module in such an application a loosely coupled module (LCM). A loosely coupled application can be defined by the combination of dependent or independent LCMs. By the loosely coupled application model, Grid applications can be built by employing discrete and heterogeneous resources on the Internet. The loosely coupled relationships among different LCMs can guarantee the robustness of the application. Parameters are defined in the application model so that application schedulers in the Grid environment can efficiently implement application scheduling by designing appropriate scheduling algorithms based on these parameters. [ABSTRACT FROM AUTHOR]

The grid computing paradigm involves both the availability of abundant computing resources, and the storage of increased amounts of valuable data. Such information systems heavily rely upon the provision of adequate security. It is imperative that techniques be developed to assure the trustworthiness of these environments. Formal verification provides the tools and techniques to assess whether systems are indeed trustworthy, and is an established approach for security assurance. When using formal verification for security assessment one of the most important concerns should be to be precise about the threat model. A comprehensive threat model is indispensable for the simulations of a grid security model. This article presents a survey of the various threat models and discusses how and when these threat models may be inappropriate for use in the grid computing environments. Then a fine-grained threat model for grid computing is presented. [ABSTRACT FROM AUTHOR]

This paper examines the problem of adapting parallel applications on a cluster of workstations. The cluster is assumed to be a heterogeneous, multi-user computing environment so that efficient load balancing within the application must take external factors into account. At any time the users of the network are competing for resources. Performance of a particular processor, as a component in the parallel (message passing) computation, depends on both static factors, such as the processor hardware, and dynamic factors, such as the system load and the activities of other users. For each processor, the external factors can be condensed into a single parameter, the load index, which is a normalised measure of the current spare capacity of the processor available to the application. Numerical experiments show the efficiency of the load balancing strategies on a finite element application with a domain decomposition and the effect on overall computation time. [ABSTRACT FROM AUTHOR]

This paper describes a workflow management system that was implemented according to the requirements of a Flood forecasting application [1] developed in the CrossGrid project [2]. The flood forecasting application contains several simulation models (meteorological, hydrological, hydraulic) implemented as the grid jobs. A need for a cascade execution of these models has been a motivation for creating workflow management system capable of executing of a cascade of simulations in the CrossGrid testbed. First implementation of the workflow system has been tied to the portal user interface, but currently has been decoupled as a standalone grid service, which can be used by the two user interfaces developed in the project. [ABSTRACT FROM AUTHOR]

In this paper we study a fault tolerant model for Grid environments based on the task replication concept. The basic idea is to produce and submit to the Grid multiple replicas of a given task, given the fact that the failure probability for each one of them is known a priori. We introduce a scheme for the calculation of the number of replicas for the case of having diverse failure probabilities of each task replica and propose an efficient resource management scheme, based on fair share technique, which handles the task replicas so as to maintain in a fair way the fault tolerance in the Grid. Our study concludes with the presentation of the simulation results which validate the proposed scheme. [ABSTRACT FROM AUTHOR]

To attract users from a wide range of scientific areas it is important to provide a variety of ways to access the Lunarc resources. Providing good user interfaces for all categories of users is a key factor for a high use of resources. One way of doing this is by providing access to common applications through a web-based portal. Existing grid portals are often geared toward's providing access to the features of the middleware instead of providing interfaces to existing applications. Lunarc has adopted a more application oriented approach providing dedicated user interfaces for each application using the grid middleware as an infrastructure for the implementation. The users as application experts can extend the portal themselves using an easy to use software development kit (SDK) for plugins. [ABSTRACT FROM AUTHOR]

Grid technology attempts to support flexible, secure, coordinated information sharing among dynamic collections of individuals, institutions, and resources. The use of Grid services requires a resourceful workstation, specialized software installed locally and expert intervention. Mobile handheld devices in general do not posses enough computational and communication assets to meet the criteria for utilizing the Grid infrastructure services. We present the design of a middleware approach that aids handheld devices in this regard by wrapping the computational and resource intensive tasks in a surrogate and shifting them to a capable machine for executionThis work is supported in part by the Korea Ministry of Information and Communications' ITRC program in joint collaboration with ICU Korea.. Reduction in computational burden at the handheld device is analyzed in a test scenario. [ABSTRACT FROM AUTHOR]

Web Services Resource Framework (WSRF) is a set of specifications that represents a convergence point of the Web services and the Grid services communities. This paper presents our early experience with WSRF. A shopping cart system has been implemented with WSRF supported Globus toolkit 3.9.2 (GT3.9.2). Based on the system, the performance of the WSRF core in GT3.9.2 has also been evaluated. [ABSTRACT FROM AUTHOR]

In these years, grid computing is probably the most promising approach for building large scale and cost effective applications. However, this very popular approach needs a sophisticated software infrastructure to address several requirements. One of these requirements is the ability to sustain a predictable performance in front to the fluctuations related to the dynamic nature of a grid. In this paper we describe design and realization of a Performance Contract System, a software infrastructure that manages the computational kernel of a grid application with the aim to face such aspect of the grid computing, as well as the strategies and the experiences to integrate it in a grid-enabling component-based programming environment still under development. [ABSTRACT FROM AUTHOR]

This paper describes algorithms and software architecture of a resource broker designed in the context of MedIGrid, a medical imaging application for the management, visualization and reconstruction on grids of medical images produced by PET/SPECT medical instruments. The broker allows the discovery and selection of suitable clusters of workstations for the execution of parallel image reconstruction algorithms. The proposed algorithms and software architecture are general with respect to possible application domains and are potentially useful in different grid environments. [ABSTRACT FROM AUTHOR]

The GriddLeS (Grid Enabling Legacy Software) middleware is a novel software layer that allows previously separate legacy applications to be coupled in a software workflow over the grid without any changes to the application source code. We have previously tested GriddLeS on a number of applications, including a small atmospheric sciences workflow in which weather and climate models are coupled. In this paper we describe a number of enhancements to the previous implementation [1] [2] that improve its performance significantly. Specifically, the new implementation improves the performance when small blocks are written over high latency networks. The paper also describes a much larger atmospheric sciences workflow than previously reported, that couples multiple global climate models, regional weather models and pollution models in one virtual application. [ABSTRACT FROM AUTHOR]

This paper is about optimizing Grid application setup by allowing a user to configure a Grid application on her own PC and thereafter migrating the entire application onto the Grid where the Grid system replicates the application including the operating system instance onto the assigned nodes within the Grid. We use this strategy to develop a new Minimal Configuration Grid model (MCG). Configuration is simplified because the user does it on her own machine. Cluster administration is simplified because only a minimal software base is required: No OS nor any shared libraries need be present; they are simply migrated to each node. We have built a prototype MCG based on Intel x86 and the Xen Virtual Machine Monitor. [ABSTRACT FROM AUTHOR]

Currently, access to Grid services is limited to resourceful devices such as desktop PCs but most mobile devices (with wireless network connections) cannot access the Grid network directly because of their resource limitations. Yet, extending the potential of the Grid to a wider audience promises increase in flexible usage and productivity. In this paper we present a middleware architectureThis research work has been supported in part by the Korea Ministry of Information and Communications' ITRC program in joint collaboration with Information and Communications University, Korea. that addresses the issues of job delegation to a Grid service, support for offline processing, secure communication, interaction with heterogeneous mobile devices and presentation of results formatted in accordance with the device specification. This is achieved by outsourcing the resource intensive tasks from the mobile device to the middleware. We also demonstrate through formal modeling using Petri nets that the addition of such a middleware causes minimum overhead and the benefits obtained outweigh this overhead. [ABSTRACT FROM AUTHOR]

This paper addresses problem of the middleware for Next Generation Grid construction and presents how two important concepts such as virtualization and provisioning could satisfy some major requirements of these systems. A vision of NGG is presented and requirements set on middleware layer are analyzed. Next an enabling technology for NGG is considered . The concept of virtualization and provisioning of grid resources is explained. Finally case study illustrating N1 technology application usage is presented. [ABSTRACT FROM AUTHOR]

Monitoring systems are nowadays ubiquitous in complex environments, such as Grids. Their use is fundamental for performance evaluation, problem spotting, advanced debugging and per-use accounting. Building such systems raises challenging issues, like data gathering from Grid components, low intrusiveness, ease of use, adaptive data visualization, fault-tolerance and self-maintenance. This paper presents a new layered architecture, named Toytle, specifically designed to address these issues in the context of control Grids. All their components, from computing and network resources to complete physical processes with soft time constraints, can be monitored with Toytle. The architecture's layers, namely the distributed core, the hierarchical connections and the local monitors, have been designed to ensure scalability, high-speed sampling and efficient dealing with large data bursts. The future Toytle implementation will adapt existing tools and also create entirely new modules. Keywords: control Grids, monitoring architecture. [ABSTRACT FROM AUTHOR]

This paper presents a virtual hierarchical overlay network- VIRGO for scalable Grid computing. This virtual hierarchical overlay network is self-organizing and decentralized, with an effective lookup protocol for routing messages. It contains an n-tuple replicated virtual tree structured network that differs from DHT-based P2P networks and random unstructured networks cached by least-recently used (LRU) and minimum difference (MinD)replacement strategies. It retains the partial-match query and robust aspects of unstructured P2P and the advantage of effective routing and guaranteed searching of structured P2P. The time complexity, space complexity and message-cost of VIRGO is O(logN), where N is the total number of nodes in the network. Since LRU and MinD replacement strategies are used for caching route nodes, VIRGO is also a load-balanced network. [ABSTRACT FROM AUTHOR]

The advent of widely interconnected computing resources introduces the technologies of grid computing. A typical grid system, the cluster grid, consists of several clusters located in multiple campuses that distributed globally over the Internet. Because of the Internet infrastructure of cluster grid, the communication overhead becomes as key factor to the performance of applications on cluster grid. In this paper, we present a processor reordering technique for the communication optimizations of data parallel programs on cluster grid. The alignment of data in parallel programs is considered as example to examine the proposed techniques. Effectiveness of the processor reordering technique is to reduce the inter-cluster communication overheads and to speedup the execution of parallel applications in the underlying distributed clusters. Our preliminary analysis and experimental results of the proposed method on mapping data to logical grid nodes show improvement of communication costs and conduce to better performance of parallel programs on different hierarchical grid of cluster systems. [ABSTRACT FROM AUTHOR]

In this paper we review the central concepts of the Open Grid Service Environment, which represents an abstract service stack and was introduced to design workflow-based problem solving environments. So far, it has been used to model large-scale computational finance problems as abstract workflows with meta-components and instantiate such workflows with different components based on semantic matching. In this paper we continue by presenting two further examples from the field of computational finance, which substantiate the need for Grid technologies and take a closer look at the implementational issues and apply the concepts of Enterprise Service Busses for parallel process orchestration. The CCA (Common Component Architecture) and its convergence to Web service specifications is the key technology for integrating heterogeneous components, with which one has to deal with in the financial sector. [ABSTRACT FROM AUTHOR]

In this work we present the design and development of an adaptive task scheduling model which enables the definition and exploitation of a framework especially suitable for managing environments of intensive computing load. The framework supplies queuing mechanisms, priority-based scheduling and resources allocation strategies, load monitoring, and implements fault tolerance procedures. Buffering strategies have been used to reduce idle time for load reposition and to take advantage of I/O overlapping, increasing efficiency in the use of resources. Several tests have been performed using applications from the bioinformatics domain for which adaptive strategies have shown their ability to produce a noticeable reduction on execution time. [ABSTRACT FROM AUTHOR]

In this paper we present the data management tasks and tools used in a flood prediction application of the CROSSGRID This work is supported by EU 5FP CROSSGRID IST-2001-32243 RTD project and the Slovak Scientific Grant Agency within Research Project No. 2/3132/23. project. The application consists of a computational core - a cascade of three simulation stages, a workflow manager, two user interfaces and a data management suite. The project is based on the Grid technology, especially the Globus toolkit 2.4 and 3.2 and the EU DataGrid project. The paper begins with brief introduction to the application, its architecture and used technology. Most of the rest of the paper then describes in detail the various data management problems of the application and their solutions. The paper is concluded with a brief description of planned future work. [ABSTRACT FROM AUTHOR]

This paper presents a framework for managing jobs in the NorduGrid ARC middleware. The system introduces a layer between the user and the grid, and acts as a proxy for the user. Jobs are continuously monitored and the system reacts to changes in their status, by invoking plug-ins to handle a certain job status. Unlike other job management systems, our is run on the client side, under the control of the user. This eliminates the need for the user to share a proxy credential, which is needed to control jobs. Furthermore the system can be extended by the user, as it is designed as a framework. This gives the users the possibility to adapt and extend it, to their needs. [ABSTRACT FROM AUTHOR]

One of the biggest obstacles in the wide-spread industrial take-up of Grid technology is the existence of a large amount of legacy code programs that is not accessible as Grid Services. On top of that, Grid technology challenges the user in order to intuitively interconnect and utilize resources in a friendly environment. This paper describes how legacy code applications were transformed into Grid Services using GEMLCA providing a user-friendly high-level Grid environment for deployment, and running them through the P-GRADE Grid portal. GEMLCA enables the use of legacy code programs as Grid services without modifying the original code. Using the P-GRADE Grid portal with GEMLCA it is possible to deploy legacy code applications as Grid services and use them in the creation and execution of complex workflows. This environment is tested by deploying and executing several legacy code applications on different sites of the UK e-Science OGSA testbed. [ABSTRACT FROM AUTHOR]

GIS Grid (geographic information systems grid) is a combination of geographic information systems and grid technology. Existing non-standardized multi-sources and multi-scales data have a shortage of spatial information shared in either internal and external organizations or departments, especially in national or global applications. It is very important and difficult to store massive spatial data for GIS Grid. And efficient metadata management and data access methods should be studied. At the same time, GIS Grid should provide users with just the services and data they need, without having to install, learn, or pay for any unused functionalities. With the integration of web service and other middleware/toolkit, which support the creation of Grid infrastructures and applications, it is possible to dynamically assemble applications from GIS Grid for use in a variety of client applications. [ABSTRACT FROM AUTHOR]

This article describes the implementation of checkpointing and recovery services in a Java-based distributed platform. Our case study is suma, a distributed execution platform implemented on top of Grid services. suma has been designed for execution of Java bytecode, with additional support for parallel processing. suma middleware is built on top of commodity software and communication technologies, including Java, Corba, and Globus services. The implementation of suma that runs on top of Globus services is called suma/g. [ABSTRACT FROM AUTHOR]

Service Oriented Programming (SOP) is an emerging paradigm for grid computing and distributed applications development. SOP models an application as a composition of services, i.e. software components either local or remotely provided by third-party organizations. In SOP service information, like location and interface description, is often stored in registries (like UDDI) distributed in internet. Before using a service, applications need to know in advance its endpoint address or the location of the registry where to look for service information. In this scenario, dynamic discovery of service registry as well as inter-registry co-operation would be a very desirable feature. This paper proposes an approach for service discovery based on DNS messaging. It provides applications with DNS-based querying mechanisms to publish and discover on the network either the location of service registries or the endpoints of web services. This approach was implemented in a middleware system that uses the Multicast DNS technology and the DNS-basedService Discovery specification to provide respectively the communication infrastructure and a standard naming convention for service registries and web services advertisement and retrieving in a LAN. [ABSTRACT FROM AUTHOR]

The development and deployment of middleware and applications in a grid environment spread over many institutions is a complex challenge. The management of the development process is of key importance for the successful completion of the software components. This paper presents the measures taken and the experience gained within the EU project CrossGrid. After briefly presenting the infrastructure, the development procedures and the supporting tools are described in detail. The CrossGrid "common directory" approach allows an effective, centrally managed configuration of distributed resources. The experience of the CrossGrid integration team is discussed. [ABSTRACT FROM AUTHOR]

This paper proposes a runtime environment for dynamically changing, parallel scientific applications. This kind of applications is motivated by the LOFAR/LOIS project aiming at a multidisciplinary research platform for natural scientists and engineers. The dynamic infrastructure in turn is than mapped to Grid Services environments. [ABSTRACT FROM AUTHOR]

We describe a hierarchical dynamic load balancing protocol for Grids. The Grid consists of clusters and each cluster is represented by a coordinator. Each coordinator first attempts to balance the load in its cluster and if this fails, communicates with the other coordinators to perform transfer or reception of load. This process is repetaed periodically. We show the implementation and analyze the performance and scalability of the proposed protocol. [ABSTRACT FROM AUTHOR]

This paper is a summary of experimental results in the domain of cost estimations of data access to data stored in the secondary storage. It mainly discusses technical and development issues based on implementation prepared for EU CrossGrid project. The proposed solution is an extension of research in the field of optimization of data access for grid environment, published previously elsewhere. It is important to mention, that the presented solution might seem very simple in comparison with already available ones based on the analytical secondary storage simulation approach, but in contrast to them this method can give precise predictions even if the information about the secondary storage internal architecture is uncertain and incomplete. [ABSTRACT FROM AUTHOR]

Traditionally, distributed resource management/ scheduling systems for the Grid (e.g. Globus/ GRAM/ Condor-G) have tended to deal with coarsegrained and concrete resource types (e.g. compute nodes and disks), to be statically configured and non-extensible, and to be non-adaptive at runtime. In this paper, we present a new resource management framework that tries to overcome these limitations. The framework, which is part of our ‘Gridkit' middleware platform, uniformly accommodates an extensible set of resource types that may be both fine-grained (such as threads and TCP/IP connections), and abstract (i.e. represent application-level concepts such as matrix containers). In addition, it is highly configurable and extensible in terms of pluggable strategies, and supports flexible runtime adaptation to fluctuating application demand and resource availability. As a key contribution, the notion of tasks enables resource requirements to be expressed orthogonally to the structure of the application, allowing intuitive application-level QoS/ resource specification, highly flexible mappings of applications to available distributed infrastructures, and also facilitates autonomic adaptation. [ABSTRACT FROM AUTHOR]

Grid technology, as a highly distributed computing environment, requires an optimized access to the data resources to increase data availability. In this paper, we propose a replication technique which is based on parallel transfers from multiple sites containing replicas of the desired file. From each site, we transport in parallel only a portion of the given data source, obtaining the whole file at the end of the process. We describe the work related to the data replication; then we discuss two algorithms for striped replication optimization that aims at the minimization of the time necessary for data transfer. Finally, we present the results of the striped replication mechanism achieved by the prototype implementation of the striped replication algorithm. We compare them with the results of the standard replication tools and show interesting performance improvement. [ABSTRACT FROM AUTHOR]

The Grid Computing[1] has been paid much attention from researchers as an alternative to parallel computing for its unlimited number of potential resources available and as an easier way to build collaborative environments among multiple disciplines. However, the difficulty in establishing the environments and accessing and utilizing the resources has prevented application scientists from using Grid computing. Thus, the present study focuses on building PSE(Problem Solving Environment) which assists application researchers to easily access and utilize the Grid. The Cactus toolkit, originally developed by astrophysicists, is used as a base frame for Grid PSE. Some modules are newly developed and modified for CFD(Computational Fluid Dynamics) analysis. Simultaneously, a web portal, Grid-One portal, is built for remote monitoring/control and job migration. Cactus frame through the web portal service has been applied to various CFD problems, demonstrating that the developed PSE is valuable for large-scaled applications on the Grid. [ABSTRACT FROM AUTHOR]

Grid-Ireland, the national computational grid for Ireland, has a centrally managed core infrastructure: the installation and configuration of grid gateways at constituent sites are controlled from an Operations Centre based at Trinity College Dublin. We have developed tools to automate and simplify the deployment of Grid middleware to these sites. Virtual machine (VM) technology has performed an important role, allowing us to maximise the utilisation of server hardware and to simplify installation and management procedures. By running multiple OS instances, each on a VM, a full LCG-compatible Grid gateway can be hosted on a single computer. This has significantly reduced the hardware, installation and management investment needed to deploy a new site. In this paper, we summarise an evaluation of competing VM technologies and relate our experience with virtual machines to date. We also describe a single-computer Grid gateway based on the Xen VM system which we plan to deploy to eleven new sites in early 2005. [ABSTRACT FROM AUTHOR]

R-GMA (Relational Grid Monitoring Architecture) [1] is a grid monitoring and information system that provides a global view of data distributed across a grid system. R-GMA creates the impression of a single centralised repository of information, but in reality the information can be stored at many different locations on the grid. The Registry and Schema are key components of R-GMA. The Registry matches queries for information to data sources that provide the appropriate information. The Schema defines the tables that can be queried. Without the combined availability of these components, R-GMA ceases to operate as a useful service. This paper presents an overview of R-GMA and describes the Registry replication design and implementation. A replication algorithm for the Schema has also been designed. [ABSTRACT FROM AUTHOR]

This paper introduces a methodology to model the execution time of several computation and communication routines developed in the frame of the CrossGrid project. The purpose of the methodology is to provide performance information about some selected computational kernels when they are executed in a grid. The models are based on analytical expressions obtained from exhaustive monitorized measurements. Even though the kernels that are considered in this work include both applications dependent and general purpose, the methodology can be applied to any kind of kernel in which the most relevant part in terms of execution time is due to computations and/or communications. We focused on MPI-based communications. In addition, an interactive Graphical User Interface was developed to summarize and show the information provided by the models from different views. [ABSTRACT FROM AUTHOR]

This paper presents a new approach for certifying the correctness of program executions in hostile environments, where tasks or their results have been corrupted due to benign or malicious act. Extending previous results in the restricted context of independent tasks, we introduce a probabilistic certification that establishes whether the results of computations are correct. This probabilistic approach does not make any assumptions about the attack and certification errors are only due to unlucky random choices. Bounds associated with certification are provided for general graphs and for tasks with out-tree dependencies found in a medical image analysis application that motivated the research. [ABSTRACT FROM AUTHOR]

The global internet is rich in commodity resources but scarce in specialized resources. We argue that a grid framework can achieve better performance if it separates the management of commodity tasks from the tasks requiring specialized resources. We show that the performance of task execution on a commodity grid is the delay of entering into execution. This effectively transforms the resource allocation problem into a routing problem. We present an approach in which commodity tasks are distributed to the computation service providers by the use of a forwarding mesh based on randomized Hamilton cycles. We provide stochastically weighted algorithms for forwarding. Mathematical analysis and extensive simulations demonstrate that the approach is scalable and provides efficient task allocation on networks loaded up to 95% of their capacity. [ABSTRACT FROM AUTHOR]