[9fans] history question ("Athena's OldFiles")

[9fans] history question ("Athena's OldFiles")

[Dave Eckhardt]

According to http://cm.bell-labs.com/plan9/about.html:
 Slowly, ideas from Plan 9 are being adopted by other systems.
 [...] The dump file system has been mimicked in Athena's
 OldFiles directories or Network Appliance's .snapshot
 directories. 

I don't know much about Athena, but I believe OldFiles
has been the name for "yesterday's read-only snapshot"
in AFS since the mid 1980's.  The snapshot is mentioned,
although not by name, in Section 6.5 ("Backup") of
"Scale and Performance in a Distributed File System"
(Howard et al., ACM TOCS, Feb. 1988), which documents
AFS as of early 1987.

My vague understanding, which could be wrong, is that
in the late 80's or early 90's, there was cross-fertilization
when Athena switched from NFS to AFS, and AFS switched
from home-grown authentication to Kerberos.  If so the
appearance of OldFiles at MIT may well be after, but not
caused by, the Plan 9 WORM file server's dump tree.

Dave Eckhardt

Re: [9fans] history question ("Athena's OldFiles")

[chad brown]

In case anyone cares...

Athena started using AFS in the late 80's, and switched mostly
wholesale in the early 90's.  OldFiles comes directly from
that time.

Athena was using Kerberos authentication in AFS right from the
start -- and I believe is still using the home-grown Kerberos support,
not the stuff that CMU added to AFS shortly thereafter.

*chad

On May 12, 2006, at 11:01 PM, Dave Eckhardt wrote:

> According to http://cm.bell-labs.com/plan9/about.html:
>  Slowly, ideas from Plan 9 are being adopted by other systems.
>  [...] The dump file system has been mimicked in Athena's
>  OldFiles directories or Network Appliance's .snapshot
>  directories.
>
> I don't know much about Athena, but I believe OldFiles
> has been the name for "yesterday's read-only snapshot"
> in AFS since the mid 1980's.  The snapshot is mentioned,
> although not by name, in Section 6.5 ("Backup") of
> "Scale and Performance in a Distributed File System"
> (Howard et al., ACM TOCS, Feb. 1988), which documents
> AFS as of early 1987.
>
> My vague understanding, which could be wrong, is that
> in the late 80's or early 90's, there was cross-fertilization
> when Athena switched from NFS to AFS, and AFS switched
> from home-grown authentication to Kerberos.  If so the
> appearance of OldFiles at MIT may well be after, but not
> caused by, the Plan 9 WORM file server's dump tree.
>
> Dave Eckhardt