- From: Mark Baker <distobj@acm.org>
- Date: Thu, 5 Dec 2002 11:27:57 -0500
- To: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>
- Cc: www-ws-arch@w3.org
Hi Mike, On Thu, Dec 05, 2002 at 10:53:25AM -0500, Champion, Mike wrote: > Can we just focus on "reliable messaging" (AFAIK, a guarantee that > a SOAP message will arrive either 0 or 1 times at its destination, > and the sender will be unambiguously informed which it was), I believe our most recent discussion on this topic was here; http://lists.w3.org/Archives/Public/www-ws-arch/2002Aug/thread.html#289 You have to be really careful how you phrase the problem. It is impossible, without additional constraints, to solve the two army problem, which your wording above appears to describe. I think DavidB said it best here; http://lists.w3.org/Archives/Public/www-ws-arch/2002Aug/0297 > or > is the larger architectural question of "reliability" something > we can dig into? I think it definitely is worth it, though past attempts on my part to bring it up have always been shot down because many people just take it as given that "reliability" == "reliable messaging". I think that works on LANs, but not on the Internet. On the Internet, "reliability" == "reliable coordination". That's not just good theory, it's been common practice since the the Internet began. Many attempts at a reliable messaging system were made, and none saw much success. Yet application protocols (which are coordination protocols) all seemed to do just fine. For example, IMAP permits the reliable coordination of email synchronization, despite not being built on a reliable messaging infrastructure. MB -- Mark Baker. Ottawa, Ontario, CANADA. http://www.markbaker.ca Web architecture consulting, technical reports, evaluation & analysis
Received on Thursday, 5 December 2002 11:23:36 UTC