NFS backward compatibility

New Message Reply About this list Date view Thread view Subject view Author view Attachment view

From: Brent Callaghan (Brent.Callaghan@eng.sun.com)
Date: 05/21/98-04:08:58 PM Z


Date: Thu, 21 May 1998 14:08:58 -0700
From: Brent.Callaghan@eng.sun.com (Brent Callaghan)
Message-Id: <199805212108.OAA18192@terra.eng.sun.com>
Subject: NFS backward compatibility



Thanks to those who have contributed their opinions on NFS v4.
Please keep `em coming.  There's at least 110 people on this
list (some are sub-lists) and there's been 8 responses
in the last couple of days.



There have been several comments referring to "NFS backward
compatibility" being a limitation as to what we can do with v4.

There is no compatibility limitation.  As an RPC-based protocol
NFS versions are not required to be supersets of previous
versions.  Each version can have its own unique set of operations
and semantics.  We would expect NFS clients to speak to NFS
2, 3, and 4 servers and similarly NFS servers should support
version 2, 3, and 4 clients - however the versions are independent
protocols.

My expectation is that v4 will continue to be a simple protocol,
easy to implement and debug, and not the 100% solution to all
distributed filesystem issues.  I think I'd be safe in assuming
that it'll be based on ONC RPC and use the IETF's RPCSEC_GSS
security, utilize filehandles, idempotent read and write
operations, and simple recovery (statelessness has never been
a tenet of NFS design).  I think we can take take the goodness
of v2 and v3 and make it better in v4.

NFS is a foundation - not a limitation.

	Brent


New Message Reply About this list Date view Thread view Subject view Author view Attachment view

This archive was generated by hypermail 2.1.2 : 03/04/05-01:45:44 AM Z CST