[997] in Kerberos_V5_Development
Re: removing user-user authentication from rcp client
daemon@ATHENA.MIT.EDU (Donald T. Davis)
Mon Feb 5 22:23:56 1996
To: Marc Horowitz <marc@MIT.EDU>
Cc: krbdev@MIT.EDU, swick@x.org
In-Reply-To: Your message of "Mon, 05 Feb 1996 19:14:19 EST."
<9602060014.AA00800@opus.MIT.EDU>
Date: Mon, 05 Feb 1996 22:27:02 -0500
From: "Donald T. Davis" <don@cam.ov.com>
hiya marc;
i don't think of rcp as an important case for u2u; it's
just the one that i'm used to explaining u2u with. i
discussed rcp in my response to sam, because he asked
about rcp and about background, and because in '88, rcp
was the service, besides x, for which we wanted to create
u2u. i agree that adding u2u to gssapi would be the right
way to go; i wanted u2u support in gssapi all along.
> I think that the model of users sharing files directly, w/o a
> filesystem, is not a common one. I don't know of many people who push
> files between local disks when both are sufficiently "available" that
> u2u would work. It's much more common to export one of the disks (NFS
> being another, separate problem), or to use email to share stuff if a
> common filesystem is inconvenient.
my earlier msg addressed all of your points. the summary
version is: u2u allows people to provide services without
forcing them to have a srvtab on an insecure machine.
that the srvtab model is more popular than the u2u model
proves nothing, since the u2u model still isn't properly
available. file-sharing by email is sometimes inconvenient.
-don davis
