Re: Directory Services Work coordination proposal

John Curran <jcurran@nic.near.net> Sat, 13 November 1993 06:38 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa27303; 13 Nov 93 1:38 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa27299; 13 Nov 93 1:38 EST
Received: from haig.cs.ucl.ac.uk by CNRI.Reston.VA.US id aa05901; 13 Nov 93 1:38 EST
Received: from bells.cs.ucl.ac.uk by haig.cs.ucl.ac.uk with local SMTP id <g.01052-0@haig.cs.ucl.ac.uk>; Sat, 13 Nov 1993 06:16:24 +0000
Received: from nic.near.net by bells.cs.ucl.ac.uk with Internet SMTP id <g.12742-0@bells.cs.ucl.ac.uk>; Sat, 13 Nov 1993 06:16:09 +0000
Received: from nic.near.net by nic.near.net id aa13871; 13 Nov 93 1:15 EST
To: Masataka Ohta <mohta@necom830.cc.titech.ac.jp>
cc: Simon Spero <ses@hillary.oit.unc.edu>, ses@tipper.oit.unc.edu, leiner@nsipo.nasa.gov, Erik@necom830.cc.titech.ac.jp, Huizer@necom830.cc.titech.ac.jp, Erik.Huizer@surfnet.nl, osi-ds@cs.ucl.ac.uk, ietf-wnils@ucdavis.edu, apples@surfnet.nl
Subject: Re: Directory Services Work coordination proposal
In-reply-to: Your message of Sat, 13 Nov 1993 14:49:23 +0200. <9311130549.AA27719@necom830.cc.titech.ac.jp>
Date: Sat, 13 Nov 1993 01:15:43 -0500
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: John Curran <jcurran@nic.near.net>
Message-ID: <9311130138.aa05901@CNRI.Reston.VA.US>

--------
] From: Masataka Ohta <mohta@necom830.cc.titech.ac.jp>
] Subject: Re: Directory Services Work coordination proposal
] Date: Sat, 13 Nov 93 14:49:23 JST
] 
] I think it the issue is not necessarily searching but the selection
] of the best servers.
] 
] Selection by the client side is the issue to be solved by policy routing.
] That is, if a policy is rich enough to select the best route to some server,
] it is also good enough to select the server with the best route.

Hmm.   I agree with what you are saying, but would rather use the phrase 
"improved service architecture" as opposed to "policy routing".

Policy routing is one tool for getting end-to-end service with particular
characteristics, but there are others (such as resource reservation).
Refer to current ID draft-braden-realtime-outline-00.ps as an example.

(Also, the phrase "policy routing" is also used heavily in the routing 
 side of the IETF, and does not imply ability to know the characteristics
 of a given path to the destination.)

] But, if we urgently need it some static scheme should be developed.

Agreed...   Having directory services do "the best they can" with respect
to path considerations is worthwhile; I just would not want to see a major
investment in this area given that changes to the service architecture may
result in much more accurate methods.

/John