From: Bill Sommerfeld (sommerfeld@sun.com)
Date: 11/01/04-04:06:37 PM Z
Subject: Re: [nfsv4] Fwd: [Internet-Drafts@ietf.org: I-D ACTION:draft-mesta-nfsv4-domain-00.txt] From: Bill Sommerfeld <sommerfeld@sun.com> Message-Id: <1099346796.20312.13.camel@thunk> Date: Mon, 01 Nov 2004 17:06:37 -0500 On Mon, 2004-11-01 at 10:41, Rick Mesta wrote: > | I wonder, though, if it wouldn't be better to avoid adding yet another > | undifferentiated usage of DNS TXT resource records and instead expand > | the use of a more application specific RR. I was thinking you might use > | a new subtype of the AFSDB record. Subtype 1 and 2 apply to AFS and > | DFS, respectively, so subtype 3 could be NFSv4IDMAP. Since AFSDB > | records already apply to the application domain shared by NFSv4, namely, > | distributed file systems, this would seem logical. Further, the domain > | of AFS and DFS users is usually (though not necessarily) defined by the > | name of the AFSDB record, so the usage is even similar. I pointed the subtyping problem out to Rick when the proposal went through internal review at Sun. But subtyping AFSDB has the same (conceptual) problem as subtyping TXT. as I understand it (mostly repeating rote what I've learned from others): subtypes of any kind are bad news, because it means that clients have to filter the retrieved resource records and can't just ask for the precise type they want -- you wind up possibly needing to fail over to TCP for queries, you complicate signature validation and dynamic update, etc., etc., _______________________________________________ nfsv4 mailing list nfsv4@ietf.org https://www1.ietf.org/mailman/listinfo/nfsv4
This archive was generated by hypermail 2.1.2 : 03/04/05-02:13:33 AM Z CST