DHCPv6 failover protocol?

[Jeff Haran]

> -----Original Message-----
> From: dhcp-users-bounces at lists.isc.org 
> [mailto:dhcp-users-bounces at lists.isc.org] On Behalf Of Chuck Anderson
> Sent: Thursday, March 05, 2009 5:43 PM
> To: dhcp-users at lists.isc.org
> Subject: Re: DHCPv6 failover protocol?
> 
> On Thu, Mar 05, 2009 at 06:30:01PM -0700, Ted Lemon wrote:
> > On Mar 5, 2009, at 4:28 PM, David W. Hankins wrote:
> >> It's also quite possible we will have something very different; a
> >> stateless server algorithm so that (n+1) servers can answer
> >> consistently without needing to update each other.
> >
> > You mean DHCPv6 load balancing?   Yes, that seems more 
> likely.   Because 
> > IPv6 has address deprecation, the need for failover is 
> unlikely to ever 
> > justify the work involved in implementing it.
> 
> Does this mean long-lived TCP sessions will break when a different 
> address is leased?  SHIM6 and/or SCTP could be solutions to this 
> problem, but I don't think the world is as optimistic as I am about 
> those.  We may need a stateful failover protocol after all.

The same problem exists with RFC 2462 autoconfigured addresses. The
prefix preferred and valid lifetimes are intended to be finite to enable
nondisruptive network renumbering, yet there is no upper limit to the
lifetimes of the TCP connections used by some applications.

Jeff Haran
Brocade