Product Description
According to John Patrick, IBM’s vice-president for Internet strategies, “thenext big thing will be grid computing.”The purpose of this book will be to describe several interesting and uniqueaspects of this exciting new topic. Grid Computing is a type of parallel anddistributed system set-up that enables and encourages the sharing ofgeographically dispersed resources. In many ways, it represents theconvergence of supercomputing and web services. The book highlights manyachievements in this innovative computer science field, and it is intended to beof value to a wide spectrum of readers around the world regardless. IBM israpidly establishing itself as the global leader in the topic of Grid Computing.This book no… More >>
Grid Computing
The writing style is too wordy and reiterates too often. It would be better to have less repetition and more useful summaries at the end of each chapter.
There are several references to IBM’s leadership in grid computing, which seems unnecessary. Furthermore, there is no mention of the Plan 9 operating system which solved many of the problems that Globus is attempting to solve.
It starts to get technical around chapter 5. So if you already know you want something resembling a grid, start there. Rating: 3 / 5
The authors have written a fine book on the potential, execution and practicality of Grid or Utility Computing. It is large ( 400 pages ) and well written book, technically accurate and blends well with other industry strategies such as on demand and Autonomic Computing. The chapters on open standards are particularly strong, well thought out and presented. The book is designed well and book production, diagrams, layout is nothing short of highest quality – in short, excellent.
The prospect of true utility computing is within reach and technically feasible. The authors bring together best deployment practices, practical guidance on integrating existing resources, and applicable case studies. This book goes a long way to assisting that projection and should become a classic standard in the field.
Full kudos~! – and a doff of the hat to both authors. Rating: 5 / 5
The reason I purchased this book was a review I read in the IEEE Software journal, that recommended highly this book as a technical book that even steps the reader through an example implementation GRID service. Well, the first part of the book is the usual “GRID vision” hand-waving how GRID-based computing will realize the concept of “utility computing” where you turn on the switch, and presto, all the computing cycles you were starving for, are there, running your scientific or engineering code or whatever…
The book unfortunately is not well-written. Far too often, sentences are not syntactically correct, obfuscating the authors’ intents. The book is definitely not suited as a technical reference, because by reading it there is no way you can implement a GRID “HelloWorld” service. And even when you read Sotomayor’s tutorial on GRID services, that actually does guide you through your first GRID service using Java WS-Core, all you’ve done is figure out how to implement Web-services running on GT4. No mention of distributed computing, how to take advantage of parallelism inherent in a computation etc. etc.
So, overall, the book serves mostly as a layman’s (or manager’s) introduction into what GRID-computing wishes it will eventually be. Rating: 3 / 5
Grid computing is extensively described here as a means of providing high powered utility computing on demand. Currently, its potential is mostly unrealised. Many companies and universities have different grid implementations, as described by the authors. The universities’ main motivation is to dragoon enough computing resources for hard research problems. While in the commercial sector, computer companies like IBM want to sell on demand access as a means of entering a hopefully vast new market.
The grid approaches in the book collectively can be contrasted with p2p computing. Grid systems tend to use more diverse and powerful hardware and relatively small number of users. Think of this as the high end, while p2p is low end (e.g. the SETI desktop application). The book describes the vast amount of effort that has gone into devising grid standards and the various toolkits, most notably Globus.
A potential problem which may occur to the reader of this book is the sheer complexity of the grid approach. Its proponents argue that this is necessary complexity. But perhaps a p2p methodology might be easier to understand and use.
An analogy is with the X.400 and X.500 email and directory standards. While these are used by some companies, many have not done so. Due to the complexity and slowness. Too heavyweight. The danger for grid computing is meeting a similar fate. It may end up occupying a small high value niche, but no more. Rating: 4 / 5