Vernon Schryver wrote:
From: Brian E Carpenter <brian(_at_)hursley(_dot_)ibm(_dot_)com>
- claim much better QoS mechanisms
IPv6 doesn't. IPv6 offers *exactly* the same QOS mechanisms as IPv4,
namely IP Integrated Services and IP Differentiated Services. (There
is also the flow label field in IPv6, but there are as yet no detailed
specs of how it will be used and no false claims either).
No false claims about IPv6 QoS? Absolutely none at all? No recent
statements in this mailing list (or maybe it was end-to-end) that IPv6
QoS will be better than IPv4 Qos? No exaggerations in the trade press?
Do you read the same trade rags and IETF lists I do?
When I see myths propagated in the IETF, I try to correct them unless
someone else gets there first. For the trade rags it is really not
worth the bother. I was referring to RFCs, which is our output.
standards committee doubling of the IPv6 address from 64 to
This was very specifically to enable an adequate (64 bit) locator
component and an adequate (64 bit) identifier component in the address.
And this was based on experience with several datagram network architectures
of the past. The only realistic alternative was variable length addresses.
But since we settled this in 1994, it seems somewhat beside the point.
Yes, that's the spin I recall on the doubling of the IPng address. It
wasn't an entirely dishonest gloss, but that's true of everything almost
every committee does.
I was there. I was one of the members of the IPng directorate pushing to
copy the locator/identifier split, which is lacking in IPv4 and was
present in other datagram address formats of the same vintage. Please
don't tell me it was spin or a dishonest gloss. It was a design decision, and
one I stand by with no hesitation.
To put it all another way, do you think IPv6 is on the schedule that was
advertised 5-8 years ago, and if not, how much has it slipped?
Dead on schedule. I've always said it would take 15 years.