I guess everyone has to get a kick in on this stupid thread...

>SGI's NFS get's 20MB/sec over Hippi.  Add the "-o bds" option and you get
>80MB/sec.  Oh, did I mention that we scale?  3 processes is 190MB/sec.
>That's bytes, not bits, bucky boy.  Cray, my ass.  AFS?  Don't make me
>laugh, what a joke.  AFS is 10-out-of-10 on the bloato meter.

I think you totally misunderstand what AFS is for.  Yes, it's big and
bloated.  But it was the single-handed savior of distributed network
performance at Iowa State University (at the time, the largest Athena
site outside MIT, with 600-700 DECstations).

The whole thing was built around lots of distributed servers.
Read-only binary servers per building, serving sized subnets.  Along
with all the shared user file space, as well.  Network contention
sucked.  NFS sucked.  No big surprise.

We brought the number of NFS read-only binary servers down from the
mid 30's to FIVE AFS servers.  And performance everywhere was still
vastly improved.  Local caching is a BIG win.  Plus it gave us stuff
like automatic fail-over and replication, etc.  Moving user storage
space over vastly improved responsiveness as well.

Maybe an individual AFS server isn't able to benchmark as fast as a
highly tuned NFS server.  But the sum of the parts made AFS a
significantly better performing OVERALL solution than anything running
on NFS.

Of course, that's without mentioning all the administrative things it
made way easier, as well.  But I think you get the point.

-----------------------------------------------------------------------------
  Michael L. VanLoon                           michaelv@MindBender.serv.net
        --<  Free your mind and your machine -- NetBSD free un*x  >--
    NetBSD working ports: 386+PC, Mac 68k, Amiga, Atari 68k, HP300, Sun3,
        Sun4/4c/4m, DEC MIPS, DEC Alpha, PC532, VAX, MVME68k, arm32...
    NetBSD ports in progress: PICA, others...
-----------------------------------------------------------------------------