[arin-ppml] questions about AC decision re: 103.

Michael Richardson mcr at sandelman.ca
Tue Dec 22 22:11:58 EST 2009

Hash: SHA1

>>>>> "Terry" == Terry L Davis <Davis> writes:
    Terry> David

    Terry> I'll take a crack at this "needs issue" from a different
    Terry> angle.

    Terry> Since I'm in the edge of critical infrastructure numerous
    Terry> ways every, I guess I'd point out a few things in the way of
    Terry> re-addressing that we don't really think about:

    Terry> - When would be a good time for your local hospital to
    Terry> re-address its Emergency Room and ICU?

  This seems like IPv4 thinking to me.
  Renumbering in IPv4 is very hard.  
  Addresses are hard coded rather than discovered by name, and DHCP can
only assign end-host addresses, rather than also assign subnets to routers.

  In IPv4 nodes had one IP per interface, and no easy way to get two.
  In IPv6, nodes have multiple IPs per interface, in seperate subnets.

  Renumbering, while not very fun, occurs incrementally.

  I know a bit about Emergency Rooms and ICUs, as well as how the
hospitals track equipment (generally.. badly).... you renumber a room
when you have moved the patient out, while you are cleaning up, and
resetting all the equipment. 
  In the future, likely this equipment is connected together with Zigbee
or bluetooth using 6lowpan and/or RPL into a /64.  Likely, all that
equipment routes through a grounded router (maybe in the bed).
  (grounded means, having an uplink, not electrical ground)

  I would design this system to be a self-contained /64, mobile in the
hospital, and I would never renumber it as long as the patient is in
that bed.  To renumber, move the patient to another bed.
  (No, none of this address space is globally routed, but all of it is

    Terry> We make this crazy assumption that our business models
    Terry> function as they did before the Internet.  They DON'T!  We
    Terry> live in a 7x24 world now; our businesses, our livelihood, and
    Terry> our own personal safety depend on the Internet, 7x24 with at
    Terry> least 4 9's reliability.

    Terry> The PA concept is so broken in this context that I can find
    Terry> no way to defend it; I understand it at a technical level but
    Terry> that does not translate to the real world we live in.

  PA is not the problem --- it works for the majority of current uses.
  Religious belief in only PA is the problem.

- -- 
]       He who is tired of Weird Al is tired of life!           |  firewalls  [
]   Michael Richardson, Sandelman Software Works, Ottawa, ON    |net architect[
] mcr at sandelman.ottawa.on.ca http://www.sandelman.ottawa.on.ca/ |device driver[
   Kyoto Plus: watch the video <http://www.youtube.com/watch?v=kzx1ycLXQSE>
	               then sign the petition. 
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Finger me for keys


More information about the ARIN-PPML mailing list