- From: <jones@research.att.com>
- Date: Thu, 5 Dec 2002 13:21:07 -0500 (EST)
- To: Mike.Champion@softwareag-usa.com, edwink@collaxa.com, walden.mathews@tfn.com, www-ws-arch@w3.org
From: "Edwin Khodabakchian" <edwink@collaxa.com> To: "'Mathews, Walden'" <walden.mathews@tfn.com>, "'Champion, Mike'" <Mike.Champion@softwareag-usa.com>, <www-ws-arch@w3.org> Date: Thu, 5 Dec 2002 08:54:18 -0800 Subject: RE: "Reliable" web services for Next Big Thing? +1. Well said. We have gone through the same mistakes and learning curve. The infrastructure that certainly help by encapsulating disconnected interactions and retries but exception and timeout management has to be replicated/done at the application level for the reasons highlited by Walden. Please note that this is true even in the case where you use SOAP over JMS. We run into this problem in every customer deployment. Edwin My hope is that we can sort out those aspects of reliable messaging that benefit from standardization whether at the transport protocol, SOAP binding level (reliability feature), the SOAP module level, or generalized solutions for the application level. DavidO suggested some headers that might hit some 80/20 sweet spot, for example. I can imagine that there are aspects that are best managed by software at the application level (although even here one would hope there are reusable design patterns), but it doesn't seem useful to assume that every application must entirely reinvent reliability for itself. --mark Mark A. Jones AT&T Labs -- Strategic Standards Division Shannon Laboratory Room 2A-02 180 Park Ave. Florham Park, NJ 07932-0971 email: jones@research.att.com phone: (973) 360-8326 fax: (973) 236-6453
Received on Thursday, 5 December 2002 13:21:39 UTC