[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [RRG] Consensus? End-user networks need their own portable address space



Hi Lixia,

Yes, by "portable" I meant:

> To me a PI prefix has 2 separable meanings: globally uniqueness, and
> portability.

I tried to give a short version in my original message:

  http://psg.com/lists/rrg/2008/msg01310.html

    End-user networks need their own portable address space.

In the context of this discussion, it is assumed the space is globally
reachable - and "their own" implies the address space is globally unique.

The rest of my message beyond the short version leaves no room for doubt about
what I was suggesting.

I was discussing the solution to the need to be able to change to another ISP
without excessive risk, cost or disruption - where that solution would be
available in the time-frame I assumed the RRG was working within.

I assumed we would be making an architectural enhancement to to the current
architectures - IPv4 and then IPv6 - and so I assumed the RRG would propose a
solution to be implemented in the next 5 to 7 years.  This would mean choosing
an approach to this problem soon.  In all the years of fussing over this
problem, no solution other than portable PI space has ever been practical, and
I think it would be unrealistic to expect that a new, non-portable solution,
will turn up in the next 9 months in a substantial enough form that we could
recommend with confidence that it e applied to the Internet.

Based on my original assumptions about time-frame and IPv4, here is a more
explicit form of the short version:

  http://psg.com/lists/rrg/2008/msg01310.html

    In the timeframe available - choosing a technique by 2009-03 to
    recommend for development and then IPv4 deployment in 2012-15 -
    there is only one approach we can consider for enabling end-user
    networks to change providers without excessive risk, cost or
    disruption:

      A new form of address space which resembles current PI space
      in that it is entirely portable, unique, public and reliably reachable,
      but which is provided in a scalable way, to avoid the well
      known problems which each PI prefix causes today.

All the map-encap schemes can do this, so it is not an unrealistic requirement.


Rough consensus has been achieved that the RRG-suggested solution need not
apply to IPv4.  In that case, perhaps the time-scale for the RRG solution
being developed and applied may be much longer than I originally assumed.  In
that case many more options become available, since IPv6 has more scope for
being altered to support solutions other than map-encap, and since I think an
IPv6 solution is less urgent.

I think that before much else can be decided, it would be good to decide
whether or not an RRG solution will be developed for IPv4 (this has not yet
been ruled out) and what the time-frame is for developing and deploying the
solution(s).

With a longer time-frame and if the RRG is to seriously contemplate a
GSE-style radical revision of IPv6, perhaps some people may suggest solutions
for the changing ISP problem which are not so clearly "portable".

However, I still think that portable (and public, unique, globally reachable)
space always will be absolutely required, for IPv6 or IPv4, since there will
never be a situation where a complex network can be automatically changed to
use another prefix without excessive costs, risks and disruption.  As long as
we have packets and routers, there will be IP addresses controlling the fate
of packets, and those IP addresses will turn up in raw form (not a FQDN) in
routers, DNS zone files, server config files etc.


BTW, can you or anyone else point me to the "Handley Proposal"?


  - Robin


--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg