Re: [GLOBAL-V6] IPV4 to IPv6 migration

[Alain Durand <alain_durand@cable.comcast.com>]

On 6/3/08 2:09 AM, "Gert Doering" <gert@space.net> wrote:

> Hi,
> 
> On Mon, Jun 02, 2008 at 03:14:45PM -0700, David Conrad wrote:
>> In addition, in both IPv4 and IPv6, a renumbering event implies
>> disrupted communications (particularly since the actual IP addresses
>> are 'cached' in applications) and some folks care about that.
> 
> And this is actually where IPv6 could make renumbering a lot easier, as
> IPv6 hosts are expected to cope with multiple addresses in parallel.
> 
> So it's not "renumber *now*", but "add new prefix, wait a few days,
> deprecate old prefix, wait a few days, remove old prefix".  Which should
> be far less disruptive.

This is true at the IP stack level. Not at the application level. Many of
them who listen to a socket still bound to the address available at startup
and nether bother rechecking... And this is certainly not true for all
back-office style application, firewall, & many others who keep track of IP
addresses.

Renumbering remains hard, and techniques to deal with it have little to do
with IPv6.

  - Alain.