RE: Visibility (was Re: Introducing the Service Oriented Architec tural style, and it's constraints and properties.
"Champion, Mike" <[email protected]>
| Newsgroups | gmane.comp.web.services.ws-arch |
|---|---|
| Message-ID | <[email protected]> |
> > > -----Original Message----- > From: Mark Baker [mailto:[email protected]] > Sent: Wednesday, February 19, 2003 10:25 PM > To: Champion, Mike > Cc: [email protected] > > As Dave points out, and I did before him, it's the methods > that matter the most for the purposes of visibility, not the headers. And as I've argued on several occasions, that's the 1997-vintage model of firewalls, not necessarily the 2003 or 2010 model of firewalls. A number of people who appear to be knowledgeable about firewalls have "+1"-ed my arguments, and I haven't seen anyone but yourself push back on them. You may wish to elaborate. > > > So again, I agree that "visibility" is an important > property, but what > > powers it is *standards*, not just HTTP. > > So where are the standard Web service APIs? Huh? APIs??? FWIW, I was thinking of standards for the format of a message such as XML (duh), Xpath (to let an intermediary inspect an arbitrary section of a message, be it in a SOAP header, the SOAP body, or the payload), specific SOAP header/processing [proto]-standards such as WS-Security, WS-Routing, XACML, etc. and content specs such as XBRL, UBL, various bits of ebXML, and so on. Any standards that you can use to explain to a stupid machine what to look for in a message and how to use its value in a routing/filtering/cacheing decision enhance "visibility" IMHO.