International Workshop on Middleware for Grid Computing
Co-located with Middleware 2003, Rio de Janeiro, Brazil
Tuesday, June 17th 2003
including Texts and Slides
Authorization Policies in the GRID: Design and Implementation
K. Keahey1, V. Welch2, S. Lang1, B.Liu3 and S. Meder2
1.Argonne National Lab. ANL, Argonne, Illinois - USA
2.University of Chicago, Chicago, IL - USA
3.University of Houston, Houston, TX - USA
Abstract: In this paper we describe our work on enabling fine-grain authorization for resource usage and management. In particular we address the needs of Virtual Organizations (VOs) to enforce their own polices in addition to those of the resource owners, both in regard to resource consumption and job management. To implement this design we propose changes and extensions to the current Globus Toolkits (GT2) resource management mechanism. We describe how we can combine policies that are resource-specific, that is, determined by the resource owner, and community-wide, that is set by the VO, and how we can enable VO-wide resource management. Our system is designed to support fine-grain authorization on job startup and management, VO-wide job as well as resource allocation management. We describe the prototype we implemented to introduce these features into GT2 and the policy language that we designed to express fine-grain policies.
A Portal Framework For Building Collaborations
Jason Novotny, Michael Russell and Oliver Wehrens
Albert Einstein Institute, Germany
Abstract: Grid enabled portals are becoming increasingly popular as a platform for providing access to Grid services and resources. Unfortunately, much of the work done in portal development has led to vertically layered solutions that work for a particular project but are difficult to extend or reuse for other projects. The GridSphere portal framework seeks to address these limitations by providing a framework that will offer external developers a model for easily adding new functionality and hence increasing community collaboration. The GridLab portal will serve as an initial prototype to showcase the GridSphere framework and provide access to services being developed within the GridLab project.
of Agents for Computer and Telecommunication Network Management
Marcos Assunção, Carlos Becker Westphall and Fernando Koch
Federal University Santa Catarina UFSC, Network and Management Lab, Florianópolis SC - Brazil
Abstract: The centralized system approach for computer and telecommunication network management has been presenting scalability problems along with the growth in the number and diversity of equipments being managed. Moreover, the increase in complexity of the services being offered through the networks also contribute for adding extra workload to the management station. The amount of data that must be handled and processed by only one administration point can lead us into a situation where we won't have enough processing and storage power to carry out an efficient job. In this work we present an alternative approach by creating a highly distribute computing environment through the use of grids of autonomous agent to ana-lyze large amounts of data, which reduce the processing costs by optimizing the load distribution and re-sources utilization.
in Grid Computing with Active Services
Ricardo Leite1, Fabiano Oliveira1, Constantino Ribeiro1, Jauvane Oliveira1, Bruno Schulze1 and Edmundo.Madeira2
1.National Lab for Scientific Computing LNCC, Dep. of Computer Science, Petropolis RJ - Brazil
2.State University of Campinas UNICAMP, Institute of Computing, Campinas SP - Brazil
Abstract: In this paper we discuss architectural aspects of middleware for grid computing based on an infrastructure of distributed clusters and/or distributed services, an access portal for a demonstration project in progress, and also some security issues. We observe, in recent works, activities in the direction of open service architectures for grid services and for web services. We also see advantages in adding facilities offered by distributed object computing environments with their interoperable references and services. Open service architectures introduce the possibility of composing structures of nested services. Additionally we discuss some security issues for instance intrusion detection.
QoS Adaptation in Service-Oriented Grids
Rashid Al-ali1, Abdelhakim Hafid2, Omer F. Rana1 and David W. Walker1
1.Cardiff Univiversity, Dep. of Computer Science, Cardiff, Wales - UK
2.Telcordia Technologies, Inc., Red Bank, NJ, USA
Abstract: A class of applications utilizing Grid computing infrastructure require the simultaneous allocation of resources, such as compute servers, networks, memory and disk storage, and perhaps other specialized resources. Collaborative working and visualization is one example of such a class of applications. In this context, Quality of Service (QoS) is related to Grid services, and not just to the network connecting these services. With the emerging interest in service-oriented Grids, resources may be advertised and traded as services based on a Service Level Agreement (SLA). Such a SLA must include both general and technical specifications, such as pricing policy as well as specifications for the resources required to execute the service -- to ensure that particular QoS requirements are being met. A QoS adaptation algorithm is presented, to enable the dynamic adjustment of behavior of an application based on changes in the pre-defined SLA. The approach is particularly useful if workload or network traffic can change in unpredictable ways during an active session. The proposed QoS adaptation scheme is not only used to compensate for QoS degradation but also to optimize resource utilization (by accepting more service requests) and maximize the service provider monetary profits.
Resource Allocation for Grid Computing: A Model and Simulation
Jacek Gomoluch and Michael Schroeder
City University, London - UK
Abstract: Resource allocation is an important aspect of Grid computing. One approach uses market mechanisms to allocate resources. In this paper, we review the literature on market-based resource allocation for Grid computing classifying approaches as model- or state-based and pre-emptive or non-pre-emptive. Many of the existing market-based approaches take it for granted that markets are an improvement. We investigate under which circumstances market-based resource allocation by continuous double auctions and by the proportional share protocol, respectively, outperforms a conventional round-robin approach. To answer this question, we develop and justify a model for clients, servers, and the market and present simulation results. The factors which are studied include the amount of load in the system, different degrees of resource heterogeneity, and communication delays.
Visualization in Grid Environment
Gilson Giraldi, Jauvane Oliveira, Rodrigo Silva and Bruno Schulze
National Lab for Scientific Computing LNCC, Dep. of Computer Science, Petropolis RJ - Brazil
Abstract: In this paper we focus on the application of grid technology for fluid visualization. Specifically, we are interested on distributed visualization using grid. The focused visualization technique is particle tracing. They can be thought as the trajectory of massless virtual particles upon the simulated flow. We propose a distributed particle tracing system and its implementation for collaborative visualization environments in computational grids. We decided for Globus as a middleware and offer solutions to minimize communications between application components during computation, a bottleneck for particle tracers in distributed memory environments. Besides, external memory (Out-of-Core) polices are discussed as an alternative approach to reduce this problem.
Middleware Approach to the Provision of Grid Middleware
Geoff Coulson, Gordon Blair, Nikos Parlavantzas, Wai Kit Yeung and Wei Cai
Lancaster University, Computing Department - UK
Abstract: In the past few years significant progress has been made in the design and implementation of reflective middleware platforms i.e., platforms that, through reflection, can be flexibly configured, and run-time adapted/ reconfigured, especially in terms of nonfunctional properties like timeliness, resourcing, transactional behaviour, and security. Recently, we have initiated a project that investigates applying our previous reflective middleware work to the demanding and novel for reflective middlewarearea of Grid middleware environments. In particular, our focus is on the webservices-based approach being adopted in the Open Grid Services Architecture (OGSA) initiative. Within this approach we are focusing on communications and resource management issues (we are specifically not focusing, for example, on data-Grid or semantic-Grid aspects).
Object-Oriented Grid Middleware Leveraging Idle Computing Power of Desktop
Andrei Goldchleger, Fabio Kon, Alfredo Goldman and Marcelo Finger
University of São Paulo USP, Dep. of Computer Science, São Paulo - SP - Brazil
Abstract: Grid computing technology improves the computing experiences at organizations by e ectively integrating distributed computing resources. However, just a small fraction of currently available Grid infrastructures focuses on reutilization of existing commodity computing resources. This paper introduces InteGrade, a novel object-oriented middleware Grid infrastructure that focuses on leveraging the idle computing power of shared desktop machines. Its features include support for a broad range of parallel applications and mechanisms to assure that the owners of shared resources do not perceive any loss in the quality of service. A prototype implementation is under construction and the current version is available for download.
Asynchronous Middleware for Grid Resource Monitoring
Vivien Quema, Renaud Lachaize and Emmanuel Cecchet
INRIA Rhàone-Alpes,SARDES Project, LSR-IMAG - France
Abstract: Resource management in a Grid Computing environment raises several technical issues. The monitoring infrastructure must be scalable, extensible, flexible and configurable to support thousands of devices in a highly dynamic environment where operational conditions are constantly changing. The french VTHD project offers a Grid Computing testbed to experiment innovative solutions. We propose to address the resource monitoring challenges in this environment by combining component-based technologies with asynchronous communications. We introduce Dream (Dynamic REflective Asynchronous Middleware), a Java component-based message oriented middleware. Asynchronous communications are used to achieve the scalability-extensibility-openness objectives whereas the component technology provides the complementary flexibility and configurability features. We think that this infrastructure is a viable approach to build a resource monitoring infrastructure for VTHD. Moreover, Dream makes the monitoring logic accessible from J2EE application servers. This allows the monitoring information to be presented as a Grid service and an easy integration within the Open Grid Software Architecture.
Management Framework For Interactive Grids
Raj Kumar, Vanish Talwar and Sujoy Basu
Hewlett-Packard Labs - USA
Abstract: Traditional use of Grid Computing Systems has been for batch jobs in the scientific and academic computing. We envision the next generation Grid computing systems to support graphical interactive sessions. In this paper, we propose a resource management framework for supporting graphical interactive sessions in a Grid computing system. We describe the high level architectural resource management framework distributed among the submission nodes, central scheduler node, execution nodes, and aggregation nodes. We then describe in detail the resource management framework on the execution node. The description of the resource management framework on the scheduler node is kept at a high level in this paper. The framework on execution nodes consists of resource management agents, an admission control system and application predictor system. The agents on the execution node are startup agents, sensor agents, monitoring agents, aggregator agents, enforcement agents, registration agents. The session admission control system is responsible for determining if a graphical interactive session can admit a new application session. An application predictor system is responsible for predicting the resource utilization behavior of applications based on data obtained from the resource management agents. The proposed framework allows for implementation of a scalable and extensible middleware for interactive grid resource management. It supports fine grained performance guarantees specified in service level agreements and brings forth some important and novel contributions to enable graphical interactive sessions on Grids.
Multimedia Collaboration System
Geoffrey Fox1, Wenjun Wu2, Ahmet Uyar3, Hasan Bulut2 and Shrideep Pallickara2
1.Indiana University, Community Grids Lab - USA
2.Indiana University, Pervasive Technology Lab - USA
3.Syracuse University, Dep. of Electrical Engineering and Computer Science - USA
Abstract: In order to build an integrated collaboration system over heterogeneous collaboration technologies, we propose a Global Multimedia Collaboration System (Global-MMCS) based on XGSP A/V Web-Services framework. This system can integrate multiple A/V services, and support various collaboration clients and communities. Now the prototype is being developed and deployed across many universities in USA and China.
Architecture for Computational Grids Based on Proxy Servers
Paulo Capellotto Costa, Sérgio Donizetti Zorzo and Helio Guardia
São Carlos Federal University UFSCar, Dep. of Computer Science, São Carlos SP - Brazil
Abstract: The Computational Grid concept allows for large-scale resource sharing. The implementation of Grid environments relies on the existence of a communications infrastructure and efficient resource management. Access control and secure communication are also required. Architectures to support Computational Grids can be developed in a variety of ways. This work presents an approach to create a Grid architecture based on Proxy Servers. This approach is intended to offer transparency and easy lightweight deployment architecture.
Towards a framework for the automatic grid enabling of MPI applications
Cristina Boeres and Vinod Rebello
Fluminense Federal University UFF, Institute of Computing, Niterói RJ - Brazil
Abstract: The Computational Grid appears to have the potential to become an important and powerful computing platform in both the scientific and commercial distributed computing communities for the execution of large-scale applications. One of the goals of the grid is to aggregate collections of shared, heterogeneous, and distributed resources to provide computational ``power'' to a parallel application. However, designing applications capable of exploiting this performance potential remains a challenge. Writing efficient programs for traditional parallel machine is not easy, for computational grids, the task is even more challenging. This paper outlines the EasyGrid methodology and a framework for the efficient and robust execution of MPI programs across clusters of Linux based computing clusters. The principal objective of this work is to identify the application-oriented middleware necessary for, as well as develop an environment for, the automatic generation of system-aware (grid-enabled) applications capable of executing in the dynamic, unstable, distributed environment offered by computational grids.
and Coordinating Grid Environments and Applications
Cristina Ururahy and Noemi Rodriguez
Pontific Catholic University PUC-RJ, Dep of Informatics, Rio de Janeiro RJ - Brazil
Abstract: The heterogeneous and dynamic nature of Grid environments place new demands on models and paradigms for parallel programming. In this work we discuss how ALua, a programming system based on a dual programming language model, can help the programmer to develop applications for this environment, monitoring the state of available resources and controlling the application so that it adapts to changes in this state. Keywords: distributed systems, parallel applications, Grid computing, interpreted languages, eventdriven communication, middleware, coordination, dynamic adaptation.
for parallel job scheduling on dynamical computer grids
Alfredo Goldman and Carlos Queiroz
University of S.Paulo USP, Dep. of Computer Science, São Paulo - SP - Brazil
Abstract: This work presents a model that allows the execution of parallel applications in a grid environment. Our main focus is on how to share the idle cycles of clusters and computers to execute parallel applications. We introduce a new model which introduces the notions of locality and adaptability. The locality is used for job allocation, and for job migration. The adaptability provides simple mechanism to allow clusters to join, or leave a grid. We also propose the middleware architecture to implement the model.