Network Distributed Computing: Fitscapes and Fallacies Paperback – Apr 22 2004
No Kindle device required. Download one of the Free Kindle apps to start reading Kindle books on your smartphone, tablet, and computer.
To get the free app, enter your e-mail address or mobile phone number.
From the Inside Flap
The problems that exist in the world today cannot be solved by the level of thinking that created them. -Albert EinsteinThe first time a baby breathes, systems designed to ease the transition from one life-support stage to the next kick in, and a mandatory cry launches the baby's lungs on their new mission: breathing air instead of "breathing" amniotic fluid. This sort of systemic state transition, fundamental to so many biological phenomena, is also critical to complex adaptive systems of other kinds, including networks of computers. It is within this shared context that I ask you to consider this volume. When I initially proposed some of the ideas for this book, my intent was to provide a balanced overview of network distributed computing (NDC) frameworks for software developers, including comparisons between them from the perspective of L. Peter Deutsch's Eight Fallacies of distributed computing. But as with most complex endeavors, the work evolved as it progressed. In particular, the notion of a fitscape became a common thread. I have Greg Doench of Prentice Hall PTR to thank for blessing the term and recognizing the strength of the metaphor as such. From a Darwinian perspective, a complex adaptive system harboring autonomous agents competing for resources can be called a "fitness landscape." As these agents become more fit for the landscape over successive generations, the landscape itself is modified in the process. I use the term "fitscape" to refer to any complex system incorporating such competitive, adaptive agents-for example, an economy at any level, a biosphere at any level, a community of interest. While the term "ecosystem" is traditionally used as a descriptive metaphor for such systems, we do often find ourselves talking about preservation of the earth's ecosystem at all costs. Hence, the implication of the word "ecosystem" seems to beg the question in that some level of stasis is implied. A fitscape, by contrast, is always in flux. It does not represent a delicate balance requiring husbandry but is rather a system in which change, based on adaptive fitness, is applauded. Software too is created and exists within fitscapes. In particular, as NDC frameworks increasingly become fitscapes in their own right, hyperproductive competition is leading to a rate of innovation that equals in each passing year the sum of all previous years. As exhilarating as it is daunting, the future of NDC remains filled with promise, despite economic and social perturbations; even the threat of war stimulates this continual exploration of novelty. I am an unapologetic proponent of technology. Sun Microsystems, my former employer, is one of few firms that has historically exhibited the sensibilities and ethics of the Network Age now upon us. As a technology evangelist for Sun from 1997 to 2002, I had the opportunity to commune with software developers from all over the world. These conversations have given me an even deeper appreciation of the difficult challenges we face in distributed computing. But they have also reinforced my otherwise optimistic nature. Despite a touch of early 21st-century ennui, my mantra has remained constant: if there's hope for humanity, it's in software. And it is equally true that if there is hope for software, it's in our humanity. An examination of the past, present, and future of NDC illuminates many possible directions in which we might progress as we seek to fulfill this hope. By offering foundational comparisons between NDC frameworks, with a filter supplied by Deutsch and using a descriptive mechanism suggested by Stuart Kauffman, I seek to provide assistance to the software developers on whom so much depends. Admittedly, my tour of duty with Sun has left me with residual platform biases. I have, in fact, illustrated most concepts using examples from Sun's offerings. But I also believe the observations and principles contained herein transcend any specific company agenda or approach; the Church-Turing thesis applies to NDC too.
From the Back Cover
Networked distributed systems: Foundations, breakthroughs, and implications
- Building tomorrow's ubiquitous, pervasive networked computing systems
- Technologies, protocols, messaging, software, integration, collaboration, security, and more
- Avoiding the eight classic fallacies of distributed computing
- The role of XML, Web services, Spaces, Jini, and other key technologies
- Ten powerful megatrends driven by networked distributed computing
Networked distributed computing (NDC) systems are driving an ongoing technological revolution that has already spawned the Internet and will soon transform the world into one ubiquitous, pervasive "information field." In Network Distributed Computing: Fitscapes and Fallacies, Max K. Goff reviews the field's crucial challenges, state-of-the-art solutions, and breathtaking future. Goff covers both the "trees" and the "forest"-showing how NDC has evolved, where it's headed, and above all, what it all means.
- Building NDC "fitscapes": new frameworks that turbocharge innovation
- Leveraging Moore's Law, Gilder's Law, Metcalfe's Law, and the latest R and D advances
- Overcoming the eight classic fallacies of distributed computing
- Enhancing collaboration, security, and dependability in networked computing environments
- Integrating wired and wireless networks: key software challenges
- Messaging and communications protocols for distributed, interoperable systems
- The roles of XML, Web services, Spaces, Jini, and other key technologies
- NDC-driven megatrends: Semantic Web, global transparency, nanotech, robotics, and beyond
Top Customer Reviews
Fitscapes and fallacies might have been better titled "The Nature and Nurture of the Internet." Fitscape is Max's word to describe the nature of distributed computing. Charles Darwin wrote about "autonomous agents competing for resources" and Max draws the conclusion that you can apply the same description to processes running in a grid of networked computers. The book begins with an explanation of the attributes of a fitscape and then talks about the protocols that are used to build a distributed application. Fitscapes and Fallaces then compares a number of fitscapes - yes, there are more than one - using Deutsche's 8 fallacies of network computing.
While the work is slightly academic, it delivers messages I wish every beginning software developer, QA technician, and IT manager would learn.
The chapter breakout:
Fitscapes and Fallacies; Ten Technology Trends; The Scope of NDC; NDC Theory; NDC Protocols; NDC Messaging; NDC Datacom: Wireless and Integration; Today's NDC Frameworks; Tomorrow's NDC Framework Options; Fallacies and Frameworks; Composability: Real-Time, Grids, and the Rise of an NDC Meta-Architecture; Innovation and Convergence; Index
Now let me go on record right away. I deal much better with practical hands-on material. I want to see code and systems. From that perspective, this book was a let-down. You aren't going to find code you can cut and paste to build a peer-to-peer client. Having said that, this *is* an excellent book to understand exactly what types of issues are involved in conceiving, designing, and building NDC systems and architectures. In fact, I could easily see this as a textbook for a college-level course on the subject. The author stays remarkably vendor- and platform-neutral, so you won't get an overdose of *just* Sun or .Net approaches.
Not having spent much time thinking about all that goes into NDC systems, I was amazed at the complexity and issues that come into play for even the simpliest applications. With the increasing significance of distributed computing, I think most IT professionals could benefit by reading through the material and becoming aware of trends that will be here in very short order.
An important enabling technology for this future is JXTA, for mobile p2p networks. He suggests that it may have greater impact than Jini. In fact, he does a point-by-point comparison of the two, to Jini's disadvantage.
While the book is put out under Sun's rubric, and Sun has the copyright on the book, the above comparison suggests a commendable objectivity. JXTA is derived from Sun's Java, but it is now independent, whereas Jini is wholly owned by Sun. Now it should be said that other sections of the book describe quite favourably Sun's products; most notably Java. But since Sun's founding, it has indeed pushed the industry envelope on networking.
A very interesting side note he made was in comparing Java with C#/.NET. As you may know, a strongly claimed advantage for the latter is that you can write in other languages like Pascal or Visual Basic, that have been suitably extended to handle .NET, and then combine the resultant binaries made from different languages, into one functional form. But, independently of Sun, others have written "translators" that take source code in various non-Java languages, and produce Java bytecode. It turns out that there is relatively little demand for this. Goff suggests that it is because the bytecode is optimised for Java, and Java is expressive enough that you might as well write in it.Read more ›
Look for similar items by category
- Books > Computers & Technology > Certification Central
- Books > Computers & Technology > Databases > Distributed Databases
- Books > Computers & Technology > Hardware > Microprocessors & System Design > Computer Design
- Books > Computers & Technology > Networking & Cloud Computing > Data in the Enterprise > Client-Server Systems
- Books > Computers & Technology > Networking & Cloud Computing > Networks, Protocols & APIs > Networks
- Books > Qualifying Textbooks - Fall 2007 > Computers & Internet
- Books > Textbooks > Computer Science & Information Systems > Networking