S Woodside wrote:
...
44 bits are provider independent and encode GPS coordinates.
It's enough to have accuracy to roughly 10 meter squares.
The 04 version reduced that to 6.4 m, but your point about it being
somewhat course is valid. I stopped at 44 bits because the goal was
creating a /48. If that is expanded to a /52 you are looking at > .5 m
resolution. While this is possible to encode, the publicly available
gear is not capable of measuring that accurately.
BUT ... that doesn't consider altitude.
Depending on the intended use, that can be further encoded in the
remaining bits to create a unique /64 per cubic .5 m, 2 km deep.
In addition, it's quite easy to see that 10
meter squares may be a little bit too big for some purposes, even if
you're just talking about individual homes. Also, it doesn't
provide a
mechanism to define an area rather than a point. I could go on.
So it's clear at least to me that under this scheme there is
not enough
space in IPv6 to do proper justice to georouting.
We have the option of using longer than /64 prefixes if necessary. I
would prefer not to go there, but given a unique need it could be done.
Tony
simon
[1] http://www.ietf.org/internet-drafts/draft-hain-ipv6-pi-addr-04.txt
--
www.simonwoodside.com -- 99% Devil, 1% Angel