RE: [security-services] ECP profile question

From
Thomas Wisniewski <>
Date
2006-02-02T22:04:58+00:00
ID
Thread
RE: [security-services] ECP profile question
Title: RE: [security-services] ECP profile question

Brian, since there is a possibility that the responseURL may be relative (and this would not be obvious from the proposed change), how about changing line 964 to just:

/ecp_assertion_consumer

Tom.

> -----Original Message-----

> From: Brian Campbell [mailto:] 

> Sent: Thursday, February 02, 2006 4:56 PM

> To: ; 

> Cc: Scott Cantor

> Subject: RE: [security-services] ECP profile question

> 

> 

> Here's a proposed correction to replace the text in the description of

> PE35:

> 

> The example on page 29 line 964 uses a ResponseConsumerURL of

> http://identity-service.example.com/abc. Since this value must be an

> AssertionConsumerService at the SP and must match (according to the

> rules in 4.2.4.4) the value of the resonseConsumerURL, it is 

> recommended

> to change it to

> https://ServiceProvider.example.com/ecp_assertion_consumer (for

> consistency with other examples in the section and to make the example

> on that would not result in an error condition).

> 

> > -----Original Message-----

> > From: Jahan Moreh [mailto:]

> > Sent: Thursday, February 02, 2006 12:32 PM

> > To: Brian Campbell; 

> > Cc: 'Scott Cantor'

> > Subject: RE: [security-services] ECP profile question

> > 

> > I'll re-open the errata and we can discuss on next call. If you have

> > proposed corrections, please submit to the list.

> > 

> > Jahan

> > 

> > 

> > > -----Original Message-----

> > > From: Brian Campbell [mailto:]

> > > Sent: Thursday, February 02, 2006 11:14 AM

> > > To: Scott Cantor; 

> > > Subject: RE: [security-services] ECP profile question

> > >

> > >

> > > > > It seems like this example would still require the 

> ECP to send a

> > > SOAP

> > > > > fault response to the service provider.  No?

> > > >

> > > > I haven't looked closely at it, but if they don't match, it's

> wrong.

> > >

> > >

> > > I believe it is wrong so we should probably re-open that errata

> item.

> > >

> > >

> > > > > Why have the AssertionConsumerServiceURL at all?  Why not

> > > just have

> > > the

> > > > > ECP always deliver the response to the responseConsumerURL?

> > > >

> > > > The IdP is the one who knows where it's authorized to send PII

> about

> > > the

> > > > user to a given provider. The client typically is deferring this

> to

> > > the

> > > > IdP

> > > > in order to keep it minimal (but with the usual privacy costs).

> > > >

> > > > The cross-check itself is to block a MitM attack where somebody

> > > intercepts

> > > > the SP's response and redirects the ECP to tell it to send the

> > > response to

> > > > it. The IdP has the metadata and the ECP authenticates it,

> > > so it knows

> > > if

> > > > it's being told to send the response elsewhere, 

> something's wrong.

> > >

> > > Fair enough.

> > >

> > >

> ---------------------------------------------------------------------

> > > To unsubscribe from this mail list, you must leave the OASIS

> > > TC that generates this mail.  You may a link to this group

> > > and all your TCs in OASIS

> > > at:

> > > https://www.oasis-open.org/apps/org/workgroup/portal/my_workgr

> > > oups.php

> > >

> > >

> 

> 

> ---------------------------------------------------------------------

> To unsubscribe from this mail list, you must leave the OASIS TC that

> generates this mail.  You may a link to this group and all 

> your TCs in OASIS

> at:

> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgr

> oups.php 

>