Since April 1 is upon us, I thought I'd let the list know that the ECN
debate has finally been resolved, pending a few details to work out...

IPv5
====

Codename: IP "Regret"

# Header Format

-----------------------------------------------------------------
|  Ver  |       Reserved        |   Protocol   |   Hop Limit    |
-----------------------------------------------------------------
|         Payload Length        |          Flow Label           |
-----------------------------------------------------------------
|                                                               |
|                        Source Address                         |
|                                                               |
-----------------------------------------------------------------
|                                                               |
|                      Destination Address                      |
|                                                               |
-----------------------------------------------------------------

(Ver)sion (4 bits):
  Protocol version, set to 5

Reserved (12 bits):
  Reserved for fun, MBZ on open Internet

Protocol (1 byte):
  Protocol number as per IANA

Hop Limit (1 byte):
  As per IPv6

Payload Length (16 bits):
  Length of payload- fragmentation not supported

Flow Label (16 bits):
  Flow label- next to addresses for fast hashing

Source Address (8 bytes):
  64-bit source address

Destination Address (8 bytes):
  64-bit destination address

# Rationale

* provides more than adequate address space
* reduces IPv6 complexity
* reduces header size to 24 bytes, down from 40 for IPv6
* leaves 12 bits of header space for experiments and future use
* makes the Internet fun again


_______________________________________________
Ecn-sane mailing list
Ecn-sane@lists.bufferbloat.net
https://lists.bufferbloat.net/listinfo/ecn-sane