RE: geo short vs long term? [Re: Geo pros and cons]

["Tony Li" <Tony.Li@procket.com>]

Jim,

Yes, geo addressing at the continental level makes much more sense
because the connectivity within a continent is pretty much a 
given.  Not so at the metro or lower layers, so it doesn't
recurse down nicely.

Note that a hybrid system that was geo at the top and provider
based below was proposed some 10 years ago, but some folks
shouted it down because they didn't feel that the aggregation
at the continental level would have a great effect.  Of course,
we had very few ISPs at that exact point in time and they
were correct.  For the moment.

I would have zero issues with continental aggregation tied
to lower layers that were topological.  [Well, ok, I don't think
that Antartica should get as much as China, but that's a nit. ;-) ]

Tony


|    -----Original Message-----
|    From: Bound, Jim [mailto:Jim.Bound@hp.com] 
|    Sent: Thursday, April 03, 2003 6:05 AM
|    To: J. Noel Chiappa; multi6@ops.ietf.org
|    Subject: RE: geo short vs long term? [Re: Geo pros and cons]
|    
|    
|    This is logical and makes sense.
|    
|    But practically, if say the U.S., Mexico, and Canada all 
|    had a toplevel
|    GEO prefix would the connectivity issue still be real or just
|    theoretical?
|    
|    Within this GEO if all had 2ffe then another set of bits 
|    for Mexico,
|    U.S. and Canada how would that exacerbate routing problems 
|    in Austrailia
|    or France?
|    
|    Thanks
|    /jim
|    
|     
|    
|    
|    > -----Original Message-----
|    > From: J. Noel Chiappa [mailto:jnc@ginger.lcs.mit.edu] 
|    > Sent: Thursday, April 03, 2003 8:35 AM
|    > To: multi6@ops.ietf.org
|    > Cc: jnc@ginger.lcs.mit.edu
|    > Subject: Re: geo short vs long term? [Re: Geo pros and cons]
|    > 
|    > 
|    >     > From: "Michael H. Lambert" <lambert@psc.edu>
|    > 
|    >     >> Geographic addressing has been discussed extensively 
|    > about 17 times in
|    >     >> the IETF, and every time it has been rejected. 
|    > Discussing it one more
|    >     >> time is not going to change this. There is *never* 
|    > going to be a rough
|    >     >> consensus *in favour of* geographic addressing. There 
|    > will *always* be
|    >     >> a lot of people against it
|    > 
|    >     > Since you have more historical context than many of us 
|    > on the list,
|    >     > could you please summarize the arguments which have 
|    > been used against
|    >     > geo? I suspect that they are mainly financial and 
|    > political, but there
|    >     > might be technical issues I'm unaware of.
|    > 
|    > Actually, I don't think that's quite correct: the argument 
|    > against it is in fact rooted in technical issues, although 
|    > there are non-technical issues in the later stages.
|    > 
|    > The analysis of geographic has been fairly well summarized in 
|    > recent discussion here on the list, but very simply, it goes:
|    > 
|    > - i) for the overhead of the routing to scale, the hierarchy 
|    > of addressing abstractions has to be reasonably closely 
|    > related to the actual interconnection topology;
|    > - ii) this means that either connectivity has to follow 
|    > addressing, or addressing has to follow connectivity;
|    > - iii) the IETF cannot mandate where connectivity gets added;
|    > - iv) connectivity gets put in where there are actual traffic 
|    > flows and/or commercial reasons to put it in.
|    > 
|    > Therefore we have to have the addressing follow the 
|    > connectivity, and an addressing scheme such as geographic, 
|    > which to scale needs to have the connectivity follow the 
|    > addressing, is not feasible.
|    > 
|    > 	Noel
|    > 
|    > 
|    > PS: Everyone, about the "this is stupid" comment - that 
|    > wasn't directed at anyone in particular, but at all of us, me 
|    > included. I suddenly had my brain start to function, and 
|    > realized that there is never going to be *rough consensus in 
|    > favour of geographic* - which is what we'd need to go forward 
|    > with it - and therefore further discussion thereof is not 
|    > useful. I hope I didn't offend anyone - if I did, my apologies.
|    > 
|    > 
|    
|