The Distributed Computing Fallacies

I've been involved with performance, networking and other similar issues since the 1970s, so rather willingly accept the truisms in The Eight Fallacies of Distributed Computing by Peter Deutsch. (The eighth one apparently was added by James Gosling. I notice that Wikipedia expresses an alternative view of the attribution of some of the others, but surely you don't really believe anything in a wiki, do you?)

Ingrid Van Den Hoogen has written (in January 2004) about them in Deutsch's Fallacies, 10 Years After and various others have commented about how universally they apply.

And now (May 2006) architect Arnon Rotem-Gal-Oz has expanded upon all of these fallacies in his weblog, and you can get it a single PDF document -- but I found that the embedded URLs in this PDF are not live, so for your convenience here are links to his individual articles (which of course have live URLs). These are at DDJ, and free registration may be required:

1. The network is reliable
2. Latency is Zero
3. Bandwidth is infinite
4. The network is secure
5. Topology doesn't change

6. There is one administrator
7. Transport Cost is Zero
8. The Network Is Homogeneous

Some good reading there, kept up to date with mentions of web services, etc. On a completely different issue, Arnon has also asked Should Architects Code? with follow-ups Should Architects Code: Round 2 and Should Architects Code: Round 3 -- plus The 7 Deadly Sins of Design

Finally, Mark Baker makes a few points in Web Services and the Eight Fallacies ... About "the nature of distributed computing on the Internet and how it is inherrently (sic) different than elsewhere" and that "HTTP defines the single most general coordination language ever developed". Read this article and see if you agree with his conclusions.