Re: (sipp) Loopback & SST

[William Allen Simpson <[e-mail]>]

> From: Steve Deering <[e-mail]>
> > One solution might be to move the SIPP loopback address somewhere else in
> > the address space.  For example, an all-nodes multicast with intra-node
> > scope sounds awfully like a loopback to me, but I'm not multicast expert.
>
> The loopback address may be used a source address, whereas a multicast
> adress may not.
>
I suggested the same thing.

I'm not sure this is a good argument against it.  When is the loopback
ever used as a source address?

Another reasonable reply was from Bob that FE prefix is technically a
local-use address which can be filtered out for misbehaving hosts.

The same thing has to be done for intra-node scope multicast
destinations anyway.  And we might as well filter all multicast sources,
too.  So, we accomplish the same thing more efficiently by using an
intra-node scope as loopback.

I still think that loopback should be FF01::1.

Actually, that reminds me of my old wish that intra-node would be scope
0 rather than 1.  Then, loopback would be FF::1, the logical successor
to 127..1.  But, this is minor.

(What's less than intra-node?  Intra-process?  Why are we reserving it?
If so, maybe we should move it up to make room for intra-cpu for
multi-cpu platforms. ;-)

[e-mail]
 ------------------------------------------------------------------------------
IETF SIPP Working Group - Archives:  parcftp.xerox.com:/pub/sipp
Unsubscribe:	unsubscribe sipp		(as message body, not subject)
Direct all administrative requests to [e-mail]