[AFS3-std] updated extended callback draft

[Matt Benjamin]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

That's a fair point, the reasoning here is incomplete, and probably
still mis-stated.

Write-on-close at the clients can never be equivalent to bounded
asynchronous writes _at the server_.

What I'm really trying to say is:  synchronous delivery of invalidates
cannot assert consistency for the system, because it presumes
consistency before the callback event--and due to (at least)
write-on-close, the presumption is in interesting cases, commonly false.

I am next making a (not fully supported) assertion that bounded,
asynchronous delivery of invalidates is sufficient for the consistency
model we _actually_ have.  That is not to say we don't have a need for a
stronger consistency model, or models.  The locking draft formalizes a
model I believe closer to posix--and provides extension mechanisms for
supporting stronger consistency models (eg, DCE, other) in future.

I'll beef this up...


Matt

Steven Jenkins wrote:
> 
> Could you elaborate on your assertion that synchronous write-on-close
> is equivalent to bounded asynchronous writes?  Specifically, could you
> provide information on the write-on-close vs write-through (or other
> semantic models) that you are assuming for 'asynchronous callback
> events'?


- --

Matt Benjamin

The Linux Box
206 South Fifth Ave. Suite 150
Ann Arbor, MI  48104

http://linuxbox.com

tel. 734-761-4689
fax. 734-769-8938
cel. 734-216-5309

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFJB0OkJiSUUSaRdSURCNS2AJwOAy7NxA8ZzfzWbywyrUcHTVU6DACeLjks
PSVdou7NfTyDW+1VYgIiGzA=
=pb5J
-----END PGP SIGNATURE-----