NAT64/NAT-PT update in IETF, was: Re: Important New Requirement for IPv4 Requests [re "impacting revenue"]

Wed Apr 22 21:39:09 UTC 2009

Iljitsch van Beijnum wrote:
> What would have helped here is more push in this direction.
> 

What really would help is more people who are not on NANOG pushing 
vendors to support IPv6. Even my Juniper SE has mentioned that I'm one 
of 2 people he's had seriously pushing for IPv6 features. Other vendors 
have just blown me off all together (we'll have it sometime).

> People who run networks can do a lot: believe it or not, the IETF really 
> wants input from network operators, especially in the early phases of 
> protocol development when the requirements are established.
> 

Serious input and participation means work and money. Too much for me. 
Doesn't help that when I was a wee one, mom dated someone who bragged 
about his status in the IETF and even had a pen. Sad way to be 
introduced to any organization, but I have seen similar mentalities 
regarding IETF mentioned here reinforcing my belief that arrogance is 
alive and I don't have the time and money to deal with it.

> Proprietary methods duking it out in the market place is nice for stuff 
> that happens inside one box or at least within one administrative 
> domain, but it would be a nightmare in broadband deployment where I want 
> my Windows box to talk to my Apple wifi base station and my Motorola 
> cable modem to the ISP's Cisco headend and their Extreme switches and 
> Juniper routers.

Sure, but the largest missing pieces for IPv6 are single box 
implementations. Proprietary NAT is single box. Will it break stuff? 
Probably, but when hasn't it? Corporate networks won't care. They'll 
deploy the vendor that supports it if that is what they want. 
BRAS/Aggregation is another single box solution but defines everything 
about an edge broadband network, supported by the access devices 
(switches, dslams, wireless ap/backhauls, etc). The layer 3 aware access 
devices all tend to have their own single box methods of security (DHCP 
snooping, broadcast scoping, etc, etc). I've seen quite a few systems 
that can't turn the security support off and break IPv6 because of it.

Jack