Re: The Motonori Problem

B.Kumar@cs.ucl.ac.uk Wed, 25 November 1992 22:47 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa10207; 25 Nov 92 17:47 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa10203; 25 Nov 92 17:47 EST
Received: from PIZZA.BBN.COM by CNRI.Reston.VA.US id aa07921; 25 Nov 92 17:48 EST
Received: from pizza by PIZZA.BBN.COM id aa24630; 25 Nov 92 17:40 EST
Received: from BBN.COM by PIZZA.BBN.COM id aa24623; 25 Nov 92 17:38 EST
Received: from bells.cs.ucl.ac.uk by BBN.COM id aa24617; 25 Nov 92 17:38 EST
Received: from weald.cs.ucl.ac.uk by bells.cs.ucl.ac.uk with local SMTP id <g.08808-0@bells.cs.ucl.ac.uk>; Wed, 25 Nov 1992 22:38:27 +0000
To: Martha Steenstrup <msteenst@bbn.com>
cc: murayama@angora.nc.u-tokyo.ac.jp, motonori@kuis.kyoto-u.ac.jp, sone@hitachi-cable.co.jp, idpr-wg@bbn.com
Subject: Re: The Motonori Problem
In-reply-to: Your message of "Wed, 25 Nov 92 15:31:57 EST."
Date: Wed, 25 Nov 92 22:38:20 +0000
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: B.Kumar@cs.ucl.ac.uk
Message-ID: <9211251748.aa07921@CNRI.Reston.VA.US>

Martha steenstrup writes:

>In your example, suppose that the intra-domain paths X1-A1, X1-A2, and
>X1-A3 in AD A and intra-domain paths B1-Y1, B2-Y1, and B3-Y1, each
>have associated with them a cost directly proportional to distance
>between the entry and exit virtual gateways.  This cost may be, for
>example, monetary or delay.  Moreover, suppose that distance
>differences in your picture reflect those in the real network.
>
>If both of the following hold:
>
>(1) AD A and AD B advertise the distance-related costs as part
>of the transit policies that apply to their respective domains.
>
>(2) The source policy corresponding to the host in AD X and applied to
>communication with a host in AD Y requests minimization of the
>distance-related cost over the route.
>
>then a route server in AD X will automatically generate the "correct
>route" using SPF according to the advertised costs.  Here the "correct"
>route is the one exhibiting the minimum distance-related cost.

1) How will AD X send the identity of chosen virtual gateways to AD A and AD
B? I thought IDPR route setup packet lists only transit ADs, not the
individual virtual gateways in the route. Am I wrong here?

If I am correct, then X has no mechanism to specify link selection 
between AD A and B.

2) Are n't you assuming a particular cost model above by using monetary cost
as an example?

I guess if I take subscriber-provider model of charging in which AD A charges
to AD X for providing connectivity to rest of the Internet, then AD A should 
only be concerned with monetary cost of using its immediate neighbour AD A.
In this situation, monetary costs of above links (between AD A and AD B or 
links between B and Y) have no relevance to AD X (otherwise I will be 
bothering about the cost of 200 or more inter-domain links whose existence 
I may not even be aware of for sending this particular message :-).).

Secondly, realistic monetary costs is not easily expressed as a single 
comparable parameter as we may think. It is difficult because there are 
infinite ways in which charging policies can be made for a link usage. 
Therefore, using monetray cost, IMHO, is probably not the right idea.

>The route server in AD X will always know of the existence of the
>three separate virtual gateways connecting AD A and AD B, because
>these are alsways advertised as part of IDPR transit policies.
>However, if AD A or AD B does not advertise different costs for
>exiting through different virtual gateways as part of its transit
>policies, then the route server in AD X will not be able to determine
>which route is minimum-cost, based on information learned solely
>through IDPR.  In this case, the route server would have to be able to
>obtain this cost information through some other mechanism (perhaps
>private communicate between domain administrators).
>If a source domain wants to make a route selection based upon criteria
>that are not advertised in transit policies, then it must be able to
>obtain this information through other means.

If advertised policies are modified by talking on phones, will that not
create internetwide inconsistencies of policy database amongst ADs? Imagine
if such modified database is passed from one AD to another. Where will buck
stop? 

Regards,

brijesh

Department of Computer Science 
University College London 
London WC1E 6BT(U.K )