Templin, Fred L wrote: > Joe, > > If you are skeptical of IPv6 and you are also skeptical > of new architectures like the ones being discussed in > RRG, what would you have us do?
I'm not skeptical of IPv6; I just wanted to point out a resource that
indicates its uptake.
I am skeptical of the architectures in RRG, having developed similar
ones in the X-Bone overlay system, and realizing the show-stopper was
the broadcast nature of encapsulation entry discovery about 8 years ago.
I would suggest we try to move forward with less engineering and more
social engineering. In specific, we need to abandon the either-or
approach, and embrace a concurrent approach to moving forward, driven by
apps. Here's the specific way I would like to see:
- modify apps to request both v6 and v4 addrs during DNS
- modify apps to issue TCP SYNs in both v4 and v6
servers receiving both v4 and v6 can drop the v4
clients should wait 1 extra RTT to see if both
are received, and connect via v6 if answering
I.e., push IPv6 out by pushing it into the web and other apps. I would
give up on the following issues:
- worrying about doubling DNS requests by asking for both
v4 and v6
- worrying about doubling TCP SYN traffic
- worrying about dual-stack
Overall, we need to stop over-smithing complex solutions to hide the
transition to IPv6 by over-optimizing transparent interoperation.
Joe
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Int-area mailing list [email protected] https://www1.ietf.org/mailman/listinfo/int-area
