In a nutshell, this book is aimed at graduate or postgraduate students and researchers who work with grid technologies. The editors have done a fine job of compiling its many self-contained chapters, ultimately creating a logical path that leads to a better understanding of the world of pervasive grids. The syntheses of the essential concepts, together with sound mathematical analysis and examples of real-life applications, make this work a useful reference for those who work in this area. I found it an interesting, though complex, topic.
An attractive property of grid technologies is that they can be applied to a very broad range of problems, addressing large-scale computational requirements with levels of power and flexibility that cannot be delivered by standalone architectures. The benefits of scalability and flexibility in network-enabled grids are, however, counterbalanced by the intricate level of coordination that has to exist across participating nodes. This substantial book provides a good high-level overview of this multifaceted topic.
The book’s content is organized into two parts, with a total of 16 chapters. Part 1 concentrates on grid theory and services. Part 2 covers the use of middleware in a wide range of grid computing applications and scenarios. Each chapter is authored by experts (a team of 48 authors overall) who not only provide theoretical coverage of the topic, but also make reference to the actual implementations they personally have realized. The outcome is an accurate exposure of the requirements for grid computing, the types of application that can be supported, the types of users that can be catered to, the methodologies for building and using grids, and the challenges that need be solved in order to make grids commonplace.
The main topics of Part 1 are: visualization services and videoconferencing; construction of a grid of grids; replica management services; computer and grid security; and grid-enabled workspaces. Chapter 1 describes grid visualization methodologies, and presents a flexible grid visualization kernel (GVK), which is capable of exploiting grid computational power effectively. Chapter 2 covers job scheduling techniques based on network performance monitoring; an actual estimation of data transfer times between grid nodes can lead to a more precise prediction of processing times. Chapter 3 presents a multimedia collaboration system that is capable of providing videoconferencing services across heterogeneous nodes of the grid, including cellular phones. Chapter 4 seeks to apply the open grid service architecture (OGSA) in order to integrate otherwise incompatible grid families of services, in the messaging between nodes as well as in client interfaces, such as Web portals. Chapter 5 covers replica-based data storage and retrieval, and investigates the use of peer-to-peer technologies in order to attain even larger levels of scale, while avoiding recourse to dedicated data management servers. Chapters 6 and 7 cover the fundamentals of computer and grid security, and focus on trust negotiation and multiparty joint authentication. These are essential building blocks for providing coordinated resource sharing at a given trust level, and across multi-institutional virtual organizations. Chapter 8 combines grid technologies and semantic Web technologies. This provides a unified way of managing distributed computing and data sources through personalized user workspaces.
Part 2, which also consists of eight chapters, is concerned with the study and exploitation of middleware, as well as with grid computational methodologies. Chapter 9 provides thorough coverage of workflow management systems, and identifies the technological challenges that need to be addressed in order to reach pervasive adoption of grids. Chapter 10 introduces grid interoperability requirements as a prerequisite to their widespread exploitation. Webcom-G, a grid operating system that provides interoperable access between existing middlewares with automatic task synchronization, is presented. Chapter 11 gives abstractions for middleware development. Chapters 12 and 13 illustrate popular tools, such as the Globus toolkit and the grid-enabled message passing interface MPICH-G2. Chapters 14 and 15 demonstrate how grid computing can be applied to both Monte Carlo computations and to systolic arrays for the solution of linear systems. The final chapter notes how the focus of grid research and development is shifting away from high-performance scientific computing, and toward large-scale business applications, which gives readers a glimpse of the bright future that lies ahead.