RE: [Diffserv] DHCP user class

["Gary Kenward" <[e-mail]>]

I believe the intent is to provide an additional dimension for service
differentiation
in addition to application stream related differentiation. For example, an
application
in use by user A would receive a different, perhaps improved, service over
the service
provided by the same application in use by user B. The actual
differentiation supported
would be up to the service provider.

The only issue that I can see with using the IP address is that all
applications
on the host using that IP address will be impacted. Thus, for example, if
the objective is
to give user A's video streams better service, then user A's email stream
will
likely be given relatively better service over other users email.

The only solution I can see o providing user differentiation on a per
application basis
is one based upon request/assignment (e.g. Diff Edge).

Gary Kenward

==========================================================================
Gary W. Kenward
[e-mail]
Advisor, Wireless Architecture               [phone]
Advanced Technology Labs                      ESN 395-1437
Wireless Solutions                                  FAX: [phone]
==========================================================================

As far as the laws of mathematics refer to reality, they are not certain;
and as far as they are certain, they do not refer to reality.
-- Albert Einstein

The contents of this email are Nortel Networks Confidential

> -----Original Message-----
> From: Manfredi, Albert E [mailto:[e-mail]]
> Sent: Sunday, July 09, 2000 5:14 PM
> To: 'Brian E Carpenter'; Diff Serv
> Subject: RE: [Diffserv] DHCP user class
> 
> 
> -----Original Message-----
> > From: Brian E Carpenter [mailto:[e-mail]]
> > Sent: Sunday, July 09, 2000 2:35 PM
> > To: Diff Serv
> > Subject: [Diffserv] DHCP user class
> > 
> > 
> > Diffservers might want to look at
> > http://www.ietf.org/internet-drafts/draft-ietf-dhc-userclass-08.txt
> > 
> > It's already passed last call and is on the IESG's plate. But it
> > says
> > 
> >    It is often desirable to provide different levels of service
> >    to different users of an IP network.
> >    In order for an IP network to implement this service
> >    differentiation, it needs a way to classify users. A simple
> >    solution to this is to use source IP addresses for 
> classification.
> >    Under this scheme, network administrators first configure network
> >    devices such as routers to recognize traffic from a particular
> >    source IP address (or address range) and handle it specially to
> >    meet the desired level of service. 
> > 
> > If you think this is a good/bad idea, *now* is the time to say so.
> > 
> >    Brian
> 
> Here's a little more quoted:
> 
> "This document describes a simple extension of the DHCP protocol
> that enables a DHCP server to assign IP addresses from different
> address pools depending on the type of users from which it receives
> DHCP requests. With this new extension, network administrators will
> be able to use DHCP to hand out the appropriate addresses to clients."
> 
> The reason I quoted this is that it seemed strange to assign service
> category based only on the source IP address, and that 
> extension explains
> that the address assigned by DHCP would indeed depend only on 
> the host,
> _not_ on the application_s_ the host was going to be running. The DHCP
> server obviously won't know the applications that would be 
> running until
> after the IP address is assigned.
> 
> Reading further into the I-D does not dispell this notion.
> 
> Seems like a strange idea in a time of multimedia apps? 
> Request for service
> differentiation belongs at the application layer, maybe not 
> exclusively, but
> that should be the norm.
> 
> Bert
> [e-mail]
> 
> _______________________________________________
> diffserv mailing list
> [e-mail]
> http://www1.ietf.org/mailman/listinfo/diffserv
> Archive: http://www-nrg.ee.lbl.gov/diff-serv-arch/
> 
>