IPv6 and HTTPS

Sun Apr 28 22:34:48 UTC 2013

On 4/28/13, Randy Bush <randy at psg.com> wrote:
>> Doing away with IPv4 isn't a sane short-term goal for anyone
> who wants global internet connectivity/reachability, period.

Breaking global connectivity is bad.   I  don't see networks turning off ipv4.

I would favor differentiation of network characteristics -- eg
Make IPv4 a service just for bulk transfer applications.
make IPv6  the best choice for interactive applications.

-- for example: large Cable providers getting together and agreeing to
 implement a 100ms RTT latency penalty for IPv4;  in other words,
heavy buffering of IPv4 traffic,  and heavy oversubscription
(Resulting in greater total performance throughput for data transfers
over Bittorrent or microtransport, but less perception of performance
for interactive applications).

This is probably what they already have,  just stop trying to throttle
IPv4 users,  so to encourage IPv6 adoption -- they just need to make
have some high capacity IPv6 only links, and make it an uncongested
service,  that will provide additional benefits to application
developers to favor it.

Under these conditions,  IPv6 service can be higher.   Don't give it
away for free;
the IPv6  Cable/DSL service should have twice the cost for the end
user as the IPv4 service does,  so that they feel the IPv6 service is
of value,   and  should include all the assistance to achieve the
greater performance.

The exhaustion of IPv4 address space also creates an inertia against
users switching around IPv4 providers (due to insufficient IP address
space available to accommodate build out of new infrastructure);
therefore,  content providers would be incentivized to get people
accessing their site over IPv6.

E.g.
 dedicated higher-capacity links for IPv6,  and less buffering to
minimize latency,  that  way  web sites initially get an incentive to
become IPv6-enabled destinations,  in the form of  perceived
improvements in performance;
without breaking connectivity.

etc. etc.

> folk who advocate disconnecting from ipv4 should lead by example or
> stfu.  either way, it would reduce the drivel level.

> randy

--
-JH