[2550] in Kerberos
Re: MIT V5 Kerberos and AFS
daemon@ATHENA.MIT.EDU (sommerfeld@apollo.hp.com)
Thu Feb 4 19:27:15 1993
Date: Thu, 4 Feb 93 19:09:19 -0500
From: sommerfeld@apollo.hp.com
To: Joseph_Jackson@transarc.com
Cc: kerberos@Athena.MIT.EDU, ramus@nersc.gov
In-Reply-To: Joseph_Jackson@transarc.c's message of Thu, 4 Feb 1993 17:37:08 -0500 (EST),
Date: Thu, 4 Feb 1993 17:37:08 -0500 (EST)
From: Joseph_Jackson@transarc.com
I have some questions concerning this statement.
Wasn't it actually Stanford that wrote the "aklog" and "asetkey"
programs?
No. I wrote the first version of aklog at some point in 1988 or 1989
while at MIT, as a temporary expedient hack. It was around 20-30
lines of code and had a number of things hard-coded in it, including
the principal name of "rcmd.atalanta" as the kerberos service to get a
ticket for.
It was gutted and completely rewritten by Jay Birkenbilt and a few
others before it was put into real "production" use; I believe that he
also wrote asetkey (I just used emacs to edit the binary server key
file).
There are clearly several different ways of doing "V5 support" for
AFS.
About the only one which makes sense, at least in a non-source AFS
installation, is doing some sort of 'V4 compatibility hack' where you
get a V4 ticket using a V5 TGT; otherwise, you'd have to rebuild all
your servers.
When I asked Barry exactly what they were doing, he declined to answer
on the grounds that it might be proprietary, except to say that what
he was proposing hasn't been implemented yet..
- Bill
