Hi Eric,
In general, when you see a Mutable security message, it general means
that when you sign or encrypt a message
with a Mutable security token, then that token has to be in the message. In
WSE, UsernameTokens are mutable
security tokens, and hence they have to be present in the message.
Does this help answer your question? If not, please let me know
Thanks,
Sidd [MSFT]
> The big picture: I want to use WS-Trust with a Security token service that
> issues a custom Security Token that I later will be sending to other web
[quoted text clipped - 62 lines]
> Thanks in advance,
> Eric Quist
Eric Quist - 02 Jul 2005 14:53 GMT
I have understood that UsernameTokens are mutable, but I guess that i doesn't
make sense to send the UsernameToken back to the client again. Wouldn't WSE
even try to verify the UsernameToken on the client again which definately
doesn't make sense.
Considering how little I have found on the web about this I start to think
that my origonal thougts are in the wrong direction. This is what I want to
accomplish:
1. I must support KerberosToken and UsernameToken.
2. I want to get the roles that the user belongs to (probably in a custom
token).
3. I want to be able to send a custom token with the roles in a secure
manner to an Activity Service or to an Entity Service. In this "business"
services I would be able to verify that the user belongs to a role.
4. It is Ok to distribute a public part of an X509 Certificate to all the
clients (to verify the server and for encryption use), but I would rather not
use certificates to authenticate the client.
My original idea was to implement this as follows:
1. Use WS-Trust.
2. Create a Security Infrastructure Service (the STS). This is the step that
I fail to complete in my original question for UsernameTokens.
3. The client sends a request to the STS with a KerberosToken (if the user
belongs to the domain) or with an Enrypted and signed UsernameToken.
4. The STS will verify UsernameTokens against a database.
5. The STS will pick up which roles an user belongs to from the database (no
matter how the user is authenticated).
6. The STS will send this roles in a Custom XML token (which I call
RolesToken).
It is to return this custom token that I can't figure out how to solve.
7. When the client application recives the Custom XML token, it will use it
to send to another Web Service (a business web service).
8. The business web service will use this RolesToken to find out if a user
belongs to a role or not.
Is this the way to go? And in that case, what would I have to do to be able
to return RSTR in a secure way so that my business web services can trust it?
I would be so glad if someone could post the code required (to make it work
with an UsernameToken in the request) after that I have created my custom
RolesToken in the IssueSecurityToken method.
Thanks in advance, Eric
> Hi Eric,
>
[quoted text clipped - 75 lines]
> > Thanks in advance,
> > Eric Quist