Re: [saml-dev] IndexedEndpointType

[Bradley Beddoes <beddoes@intient.com>]

Scott,

That exactly backs up my point of view on the topic so thanks the quick 
response.

regards,
Bradley

Scott Cantor wrote:
>> The discussion basically is revolving around "does a different binding
>> make an endpoint non-identical, even though it is the same overall type".
> 
> No. "Identity" is based on the profile (thus the element's QName), not the
> binding. A set of related profiles might also be equated if they choose to
> overload themselves against the same element name (i.e. SAML 1 and SAML 2
> SSO).
> 
> There are two reasons to use indexes: cluster affinity and binding
> differentiation. So one of the use cases is by definition based on the fact
> that bindings don't make endpoints "non-equivalent".
> 
>> For example can a SP with two AssertionConsumerService endpoints defined
>> in metadata use the same index for GET and POST bindings or should they
>> be unique? (within that descriptor).
> 
> They must be unique. If the element is the same as another, it gets a unique
> index.
> 
> -- Scott
> 

-- 
Bradley Beddoes
Intient Pty Ltd

http://www.intient.com
begin:vcard
fn:Bradley Beddoes
n:Beddoes;Bradley
org:Intient Pty Ltd
adr:;;;;;;Australia
email;internet:beddoes@intient.com
title:Lead Software Architect
tel;work:07 31024560
tel;cell:0413 678 802
url:http://www.intient.com
version:2.1
end:vcard