[OpenAFS] the notion of "site" is not always well-defined / "project cells"

Adam Megacz megacz@cs.berkeley.edu
Sun, 29 Jan 2006 15:19:25 -0800

[ As a side note, I finally got this whole thing working by creating the  ]
[ bogus AFSDB entry mentioned earlier.  But I'd like to continue this     ]
[ discussion because I think the use case I've set up is a very valuable  ]
[ one.  I'm currently working on a detailed guide on how to set up        ]
[ "project cells" like mine and how to give users the "shortest path to   ]
[ setup".                                                                 ]

Jeff, my main concern is that if every project took the approach you
advocate, no user would be able to join two or more projects without
understanding the details of configuring OpenAFS manually.

> You are an organization attempting to support your users.

I'm not an organization.  I'm a project that cuts across a large
number of other "primary" organizations, as well as individuals who do
not belong to any relevant organization.

I think the confusion comes from the fact that AFS was originally a
commercial software program that you had to pay a huge amount of money
for.  Therefore, every user had exactly one "site" -- the organization
that paid for his/her copy.  It was extremely rare for any person to
have their client software paid for twice by two different
organizations.  But this is no longer the case.

I argue that, in the post-Transarc era, there are a large number of
situations where OpenAFS is useful for which no coherent/meaningful
definition of "site" exists ("cell", of course, is still well-defined).

  - a

PGP/GPG: 5C9F F366 C9CF 2145 E770  B1B8 EFB1 462D A146 C380