Mike Jackson writes:
>Hallvard B Furuseth wrote:
>> Note that most clients authenticate in what is normally a more secure
>> and sensible way: With the Bind operation - so the server rather than
>> the client will check the password.  Clients which do that can't
>> authenticate against such entries, since you can't tell Bind which
>> password attribute to use.
>>
>
> The userPassword attribute is multi-valued. I'm not sure about OpenLDAP,
> but I do remember that Netscape/Fedora DS can bind against any of the
> userPassword values. As well, mail is multi-valued. If you would use
> mail as a multi-valued RDN, then given any valid RDN and any valid
> userPassword, the bind should be successful.

Yes.  I should have said, the original poster can't use Bind to achieve
his goal with that scenario, since an account owner should not be able
to authenticate with e.g. email2 + password5.

> The problem here is that within one account "user A" could spoof to be
> "user B", while using "user A's" password. For example, user A could say
> "my username is user B, and here is my user A password". I would be
> almost willing to bet that none of the users within a single account
> would ever figure this out or try this.

Could easily happen by accident: Someone with access to several of an
account's addresses didn't remember which password went with which
address.

> Still, the risk does exist.
>
> Another way to model this type of multi-user account is to create an
> object for the main account, and then create child entries for the
> sub-accounts.

Yup.

-- 
Regards,
Hallvard

---
You are currently subscribed to [email protected] as: [EMAIL PROTECTED]
To unsubscribe send email to [EMAIL PROTECTED] with the word UNSUBSCRIBE as the 
SUBJECT of the message.

Reply via email to