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

[Sean Doran <smd@ab.use.net>]

On Sunday, Apr 6, 2003, at 04:00 Europe/London, Randy Bush wrote:

> if you don't count the oceanic island nations, se asia, ...  all of
> which backhaul a long way.  net topology is extremely different
> from geography.  this repeating idea of geographic aggregation
> simply ignores reality.
Standing in my building does not give you the right to expect
me to deliver packets to you.

If your GPRS operator gives you an address related to my
building's physical location, and packets arrive via my immediate
provider to me for that address, rather than to you through your GSM 
operator,
I will drop them on the floor, even though *in principle* I could 
deliver them
to you via, say, wavelan, or a 100baseTX port.   (I don't really want 
you on my
network though; I have more business-productive things to pay for
than the porn you're surfing while in my lobby, never mind security 
issues...)

Your GPRS operator, therefore, will have to give you addresses
that are unrelated to my building's physical location, even when
you're standing in it.

If we ever have a "visitor" network, you can bet that those packets 
will cross
only the public Internet, not my corporate WAN.   You'll get public IP
addresses rather than internal (and mostly 1918) ones, because you
will only be allowed to connect to the visitor wireless LAN
(or the visitor 100baseTX plug) no matter what desk you put your laptop 
upon,
even if it's mine.

	Sean.