[OpenAFS] scalability

John Bleichert John Bleichert <syborg@stny.rr.com>
Sat, 16 Mar 2002 15:35:26 -0500 (EST)

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.
  Send mail to mime@docserver.cac.washington.edu for more info.

Content-ID: <20020316152602.U50325@picard.vonbek.dhs.org>

On Sat, 16 Mar 2002, Paul Blackburn wrote:
> There are some pretty big AFS cells around. I know of one at 10 terabytes.
> You only need 1 AFS administrator for a 100 gb cell, until (s)he goes on
> holiday. ;-)
> I would steer well clear of any "NT base solution" because of:
>     a) cost
>     b) security
>     c) reliability
> --
> cheers
> paul                                   http://acm.org/~mpb

We have probably (I'm guessing) 50+ afs cells in /afs where I work,
scattered all over the globe. I can't give you disk store stats as I dont
really know them but they're probably rather large. There are probably 200
people in our local cell, and we do *not* have any sort of dedicated admin
for it. A couple engineers handle anything that pops up as an ancillary

Some people tried to push an NT storage solution but *then* we started
gathering people who had to baby-sit the things and it effectively locked
out the Unix people as nobody felt like admin-ing samba on top of the NT
stuff (we'd need to hire somebody, nobody has the time). We're now pushin
back with AFS as the win32 AFS client works fine and doesn't lock out
anybody (afs clients exist for Unix, win32 and OSX).

Obviously, the workstations will always be Unix/AFS. The win32 involvement
is an externality from choosing win32-only groupware, but that's a
separate and very bitter rant.

Bottom line, we're a 200+ cell in a planetary network that's always up,
and we don't really even have an admin. Just AFS on reliable servers.


> David Bear wrote:
> > I'm wondering if anyone has any metrics or 'gut feeling' for how scalable
> > AFS is on 2 sides.  Side one is technically, how many clients, how much
> > space, etc.  Side 2 is administratively.  I'm interested in comparatives
> > as well: ie how many sys admins does it take to manage 100gigs of AFS
> > space  and 40000 users, versus some NT base solution.  What kind of
> > incremental price points are there?  meaning, 1 afs sysadmin can handle x
> > gigs and y users, 2 afs sysadmin can handle n*x gigs and m*y users -- etc,
> > versus doing the same thing for an NT based network.
> >
> > Thanks for any input.
> >
> > --
> > David Bear
> > College of Public Programs/ASU
> > 480-965-8257
> > ...the way is like water, going where nobody wants it to go
> >
> > _______________________________________________
> > OpenAFS-info mailing list
> > OpenAFS-info@openafs.org
> > https://lists.openafs.org/mailman/listinfo/openafs-info
> --
> cheers
> paul                           http://acm.org/~mpb

|-John Bleichert----syborg@stny.rr.com----------------|

Content-Type: TEXT/X-VCARD; CHARSET=us-ascii; NAME="mpb.vcf"
Content-ID: <20020316152602.P50325@picard.vonbek.dhs.org>
Content-Description: Card for Paul Blackburn
Content-Disposition: ATTACHMENT; FILENAME="mpb.vcf"

org:IBM Managed Security Services;EMEA Security CoC Team
fn:Paul Blackburn