SSO authentication: a new approach
rplante at ncsa.uiuc.edu
Mon Mar 14 21:00:08 PST 2005
On Mon, 14 Mar 2005, John Good wrote:
> I can't see that I would be willing to let
> someone with one of your "weak certificates"
> do much more than someone with an HTTP cookie.
> I would not, for instance, let them have file
> upload access (unless I wanted to be in the
> business of supplying free storage to the
The motivation for this is the fact that many sites happily do this now,
albeit in restricted ways.
Take the example of the what we want to do with the OpenSkyQuery Portal.
We want users to be able to upload their own catalogs as a VOTable
(perhaps something they've constructed themselves using client tools or
was the result of a query from another site) in order to cross-correlate
it with a skynode. Now consider how we would do this without certs and
such: we set up a little registration page, a user database, and use
cookies to track the session. There is nothing preventing a user from
entering a false identity, but who cares? We will place restrictions on
how big a file, maybe its contents, etc. All we care about is if we allow
this uploaded data to persist that only the person who registered can get
access to it, whoever that schmo is. This is common practice all over the
web; we all have accounts like this (travelocity, blogs, etc.).
What a cert-based model can give us is interoperability between portals
offering storage space for what ever reason. A weak cert is
equivalent to the scenario above. Some sites will be more general
purpose in terms of what allow users to store, and thus may want a strong
cert that guarantees identity, but some can be more restrictive and not
worry about true identities. But we can have a common authentication
model based on certs.
The number one user complaint about certificates is that is such a pain in
the ars to get one. And as many sites maintaining persistant state today
demonstrate, they are not needed. Weak certs allow users to do simple
More information about the grid