Grid Computing

Infrastructure, Service, and Applications

Lizhe Wang, Wei Jie, Jinjun Chen

• Publisher: CRC Press
• Year: 2018
• ISBN:
  9781351834742
• Edition: 1
• Topic: Informatica
• Subtopic: Ingegneria informatica

Identifies Recent Technological Developments Worldwide

The field of grid computing has made rapid progress in the past few years, evolving and developing in almost all areas, including concepts, philosophy, methodology, and usages. Grid Computing: Infrastructure, Service, and Applications reflects the recent advances in this field, covering the research aspects that involve infrastructure, middleware, architecture, services, and applications.

Grid Systems Across the Globe

The first section of the book focuses on infrastructure and middleware and presents several national and international grid systems. The text highlights China Research and Development environment Over Wide-area Network (CROWN), several ongoing cyberinfrastructure efforts in New York State, and Enabling Grids for E-sciencE (EGEE), which is co-funded by the European Commission and the world's largest multidisciplinary grid infrastructure today.

The second part of the book discusses recent grid service advances. The authors examine the UK National Grid Service (NGS), the concept of resource allocation in a grid environment, OMIIBPEL, and the possibility of treating scientific workflow issues using techniques from the data stream community. The book describes an SLA model, reviews portal and workflow technologies, presents an overview of PKIs and their limitations, and introduces PIndex, a peer-to-peer model for grid information services.

New Projects and Initiatives

The third section includes an analysis of innovative grid applications. Topics covered include the WISDOM initiative, incorporating flow-level networking models into grid simulators, system-level virtualization, grid usage in the high-energy physics environment in the LHC project, and the Service Oriented HLA RTI (SOHR) framework.

With a comprehensive summary of past advances, this text is a window into the future of this nascent technology, forging a path for the next generation of cyberinfrastructure developers.

Part I

Grid Infrastructure and Middleware

1

CROWN: A Service Grid Middleware for e-Science

Chunming Hu and Jinpeng Huai

CONTENTS

1.1 Background

1.2 CROWN Middleware

1.2.1 Design Motivation

1.2.2 Architecture

1.2.3 Components in CROWN

1.2.3.1 Node Server

1.2.3.2 Resource Locating and Description Service (RLDS)

1.2.3.3 CROWN Scheduler

1.2.3.4 CROWN CommSec

1.2.3.5 CROWN AuthzService

1.2.3.6 CROWN CredMan

1.2.3.7 CROWN CredFed

1.2.3.8 CROWN ATN

1.2.3.9 CROWN Portal and Rich Client Framework

1.2.3.10 CROWN Designer

1.2.3.11 CROWN Monitor

1.2.4 Main Features

1.2.4.1 Overlay-Based Distributed Resource Management

1.2.4.2 Remote and Hot Deployment with Trust (ROST)

1.2.4.3 JSDL-Based Job Submission and BES-Based Job Scheduling

1.2.4.4 Security Architecture Supporting Domain Interoperability

1.3 Resource Management in CROWN

1.3.1 Overview

1.3.2 Node Server

1.3.3 S-Club and RCT

1.4 Security Architecture in CROWN

1.4.1 Overview of CROWN Security

1.4.1.1 Node-Level Security

1.4.1.2 Domain-Level Security

1.4.1.3 Region-Level Security

1.4.2 Design and Implementation

1.4.2.1 Security Structure for the CROWN Node

1.4.2.2 Communication Security

1.4.2.3 Policy-Based Authorization

1.4.2.4 Credential Management

1.4.2.5 Trust Management and Negotiation

1.4.2.6 Credential Federation

1.4.3 CROWN Security Summary

1.5 Testbed and Applications of CROWN

Acknowledgments

References

1.1 Background

In the year 2004, with the development of grid technologies in both academia and industry, the Natural Science Foundation Committee of China (NSFC), which is one of the main funding parties in China, announced its e-Science Program named Network-based e-Science Environment. As one of the grid-related research programs, this program is also referred to as the NSFCGrid. The program started in 2004 and was completed at the end of 2007. The main goal of this program is to build up a virtual scientific and experimental environment to enable wide-area research such as large-scale computing and distributed data processing. The research projects are organized into three layers: basic theory and principles, general testbed, and pilot applications. In contrast to CNGrid and ChinaGrid, NSFGrid pays more attention to the fundamental research of grid-related technologies.

CROWN is the brief name for China Research and Development environment Over Wide-area Network [1], one of the main projects in the NSFCGrid program, which is led by Beihang University (BUAA), with several partners in China, such as the National University of Defense Technology of China (NUDT), Peking University, the Computer Network Information Center of China Academy of Science (CNIC, CAS), and Tsinghua University. The output of the CROWN project falls into three parts: the middleware, set to build a service-oriented grid system; the testbed to enable the evaluation and verification of grid-related technologies; and the applications.

The CROWN service grid middleware is the kernel for building an application service grid. The basic features of CROWN are listed as follows: first, it adopts an Open Grid Services Architecture/Web Services Resource Framework (OGSA/WSRF) compatible architecture [2]; second, considering the application requirements and the limitations of the security architecture of OGSA/WSRF, more focus is put on the grid resource management and dynamic management mechanism in the design stage, and a new security architecture with a distributed access control mechanism and trust management mechanism, which are proposed to support the sharing and collaborating of resources in a loosely coupled environment, is proposed in CROWN.

Under the framework of the CROWN project, we also created the CROWN testbed, integrating 41 high-performance servers or clusters distributed among 11 institutes in 5 cities (as of April 2007). They are logically arranged in 16 domains of 5 regions by using the CROWN middleware. The testing environment is growing continuously and has become very similar to the real production environment. The CROWN testbed will eventually evolve into a wide-area grid environment for both research and production.

CROWN has become one of the important e-Science infrastructures in China. We have developed and deployed a series of applications from different disciplines, which include the Advanced Regional Eta-coordinate Numerical Prediction Model (AREM), the Massive Multimedia Data Processing Platform (MDP), gViz for visualizing the temperature field of blood flow [3], the Scientific Data Grid (SDG), and the Digital Sky Survey Retrieval (DSSR) for virtual observation. These applications are used as test cases to verify the technologies in CROWN.

1.2 CROWN Middleware

1.2.1 Design Motivation

Since many researchers are focusing on technologies and applications of grids, the interoperability and loosely coupled integration problem between different grid systems are now becoming a hot topic. At the same time, the application and standardization of Web services technology are developing rapidly, and service-oriented architectures (SOAs) have become an important trend in building a distributed computing environment for wide-area networks, which helps the merging of grid and Web services. Recently, the OGSA [2] and WSRF were proposed and have become two of the fundamental technologies in grid competing. SOA and related standardization work provide an important methodology to the research and application of grid technology. First, the resources are encapsulated into services with standardized interfaces, supporting the unified service management protocol, which helps to solve the problem caused by the heterogeneity of resources; second, the resources are utilized through a service discovery and dynamic binding procedure, which helps to set up a loosely coupled computing environment. But the current resource management mechanism is not enough for all the grid application scenarios because of the distributed and autonomic resource environment, an...