zero suppression vs compression in addresses

Bjørn Mork bjorn at
Mon May 18 12:44:24 CEST 2009

<michael.dillon at> writes:

> So, this draft *DOES* change what is considered to be
> a syntactically correct when it allows "::" to be used
> twice. 

That's clearly not the intent.  For reference, quoting from the
beginning of section 4:

  "A recommendation for a canonical text representation format of IPv6
   addresses is presented in this section.  The recommendation in this
   document is one that, complies fully with [RFC4291], is implemented
   by various operating systems, and is human friendly. "

> Admittedly, section 4.2.3 is not very clear and it may not survive
> future discussion,

I guess it should not survive, given your interpretation of it.

The point of section 4.2.3, as I read it, is: If there are more than one
sequence of 16 bit groups of zeroes, compress the longest.  In case of a
tie, compress the first sequence.


2001:db8:0:0:0:1:0:0 => 2001:db8::1:0:0 (longest match win)
2001:db8:0:0:1:0:0:0 => 2001:db8:0:0:1:: (longest match win)
2001:db8:0:0:1:0:0:1 => 2001:db8::1:0:0:1 (first of equal length win)


More information about the ipv6-ops mailing list