--- gert@space.net wrote: From: Gert Doering gert@space.net On Sun, Oct 29, 2017 at 12:57:54PM -0700, Scott Weeks wrote:
Yeah, it's quite unfortunate that IPv4 ran out so suddenly, barely 15 years after people were told to move towards IPv6.
Especially after IETF made it backwards compatible and made it so easy to switch from 4 to 6... ;-)
There's no way to make "something with longer addresses" compatible without IPv4 without changing everything (routers, endpoints) - so, that argument is usually one brought forward as one of a long list of standard excuses to avoid deploying IPv6, while at the same time blaming everyone else for the problems with IPv4. -------------------------------------------
Note the smiley face above. This one, too... :)
I was not around for those discussions (and not being a computer science person, nor wanting to go on this for too long as has been endlessly done on other lists), but it seems TLV would have allowed 4 to be a subset of the new space. I never heard that discussed much and that's what I meant by my comment.
scott
Hi,
On Sun, Oct 29, 2017 at 02:53:41PM -0700, Scott Weeks wrote:
I was not around for those discussions (and not being a computer science person, nor wanting to go on this for too long as has been endlessly done on other lists), but it seems TLV would have allowed 4 to be a subset of the new space. I never heard that discussed much and that's what I meant by my comment.
The point is: if you introduce a change to the packet format (and TLV would be), you are no longer compatible with IPv4. Which makes the whole "I want this to be compatible so I do not have to change infra or end points" totally moot.
Worse, then you have "old IPv4" and "new IPv4" machines who might or might not be able to talk to each other, depending on which IPv4 address the "new IPv4" got (a long one or a short one) - while with IPv6, you have unmodified old IPv4 to ensure compatibility during the transition, and then you turn it off (in 10 years or so).
Gert Doering -- NetMaster