[ih] IETF relevance (was Memories of Flag Day?)

[John Levine]

It appears that Scott Bradner via Internet-history <sob at sobco.com> said:
>mixed picture of IETF relevance
>
>the world's telephony runs over IETF technology (SIP & RTP) except for some pockets of analog phones, 
>IETF technology specified by ITU and 3GPP

I'd say WebRTC is a significant success. That's how videoconferencing
in browsers works (and in apps which are browser skins.) 

There's also incremental stuff like TLS 1.2 and 1.3, and also QUIC,
which fixes the performance bugs of HTTP.

I don't follow IoT very closely but it is my impression that COAP and CBOR
are widely used. 

>IPv6 has had a very slow deployment - most, imo, because it does not offer enough difference from IPv4
>and because NATs have reduced the pressure to change - but, according to Google, they are getting a lot
>of IPv6 queries (about 45%)

In retrospect it was a big mistake not to have a migration path from
IPv4 to IPv6, and the IPv6 crowd did themselves no favors by insisting
for years that it was perfect and it didn't need DHCP and a bunch of
other stuff that it now finally has. People also discovered that you
can put a lot of users behind a NAT with most of them not even
noticing, and that once we gave up on the fiction that you can't sell
IP addresses, vast numbers of underused addresses came on the market.
e.g., MIT renumbered from their /8 down to a /11 and sold off the
rest. Often when a small network wonders whether to add IPv6, it's
cheaper just to buy another block of v4 and put it in front of a NAT,
so IPv4 will never die, nor is there any compelling reason for it to
do so.

On the other hand, there is a whole lot of IPv6 in use that nobody
notices. Pretty much every mobile network is IPv6 internally with some
kind of 464NAT for backward compatibility. Every large cable network I
know has native IPv6, and a whole lot of application traffic to web
servers and Netflix goes over v6 without anyone noticing. It varies a
lot by place; India is 70% IPv6, the US 55%, the UK 40%.

All that is why the IETF has no interest in the silly proposal to turn
IPv4 Class E addresses and some part of 127/8 into public addresses.
Since it's not backward compatible, it'd need some kind of flag day,
and a lot of incentives for all the routers and other infrastructure
to change. Why would Cisco and Juniper spend money on that when their
customers already know how to buy existing v4 space or move stuff to
v6?

R's,
John