"Brian Haberman" <[EMAIL PROTECTED]> writes: > Thomas Narten wrote: > > > > > If plen = 0, then isn't the "network prefix" portion 0 bits long? > > > > That would be one interpretation. Another possibly would be that the > > network prefix should have special meaning, other than looking at the > > first plen bits. I guess I'm asking if it would make sense to do so, > > for (say) debugging purposes.
> The problem is that SSM requires all 128-bits to represent the source. To be clear, you mean all 128 bits of the source address in the IP header. I'm not questioning that or suggesting otherwise. > So, it seems simpler to indicate in some manner that the destination > multicast address is SSM and refer to the source address to determine > the "owner". Agree with this also. But, does that mean it's best to require that the 64plus middle bits of the SSM address be all zeros? Seems like it would be useful to allow (or even suggest in the SHOULD sense) that the middle bits identify the link. That way if you look at the multicast address (independant of the source IP address) you can still glean some useful information out of it. Note: I'm asking this as a question. If other folks have opinions, I'd like to hear them. > > > > > The SSM range should be FF3x::/12 (and this is mentioned in the SSM > > > overview doc today). > > > > draft-holbrook-ssm-arch-02.txt talks about reserving ff3x and ff2x for > > SSM. > I believe that this draft is in the process of being revised prior to > a last call. > > > > draft-ietf-ssm-overview-01.txt says: > > > > > Source-Specific Multicast (SSM) : This is the multicast service model > > > defined in [SSM-ARCH]. An IP datagram is transmitted by a source S to > > > an SSM destination address G, and receivers can receive this datagram > > > by subscribing to channel (S,G). SSM is derived from EXPRESS [EXPRESS] > > > and supports one-to-many multicast.The address range 232/8 has been > > > assigned by IANA [IANA-ALLOC] for SSM service in IPv4. For IPv6, the > > > range FF3x::/12 is defined for SSM services [SSM-IPV6]. > > > > But that range also covers the address that are defined in this > > document, including non-SSM addresses! (And it doesn't cover the plen > > field that defines whether the address is SSM or not!!) So things do > > not seem to be in sync. > So, since I am the common denominator between the two drafts, I must > claim responsibility. The SSM overview draft should state that > FF3x::/32 is the SSM range for IPv6. Either I still don't get it, or this isn't true either. Only if plen == 0 is the address an SSM one. So you can't say that ff3x::/32 is the SSM range, because the plen field is in the last octet of that range. Thomas -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
