Re: Another proposal

[[e-mail]]

	 Here's the proposal: As in RFC 1063, the sender attaches an MTU Request
	 option to some of its outgoing datagrams.  The option is updated by
	 routers along the way.  When the last-hop router is reached (i.e.,
	 a router which can deliver the packet directly to the destination
	 host), that router now knows the path MTU.

Without commenting (for now) on the substance of this proposal, I
should note one or two glitches with it.  First, the last-hop router
does not necessarily know the host's MTU.  From the Host Requirements
RFC, it is not clear that hosts are required to accept packets of
length greater than 576, regardless of the MTU.  (The discussion, at
least as much of it as I could find from home, wasn't that clear.  It
does say explicitly that hosts *should* accept packets up to the
network's MTU after reassembly, and that hosts are only required to
accept 576 bytes.)  If the router has no need to talk directly with the
host, it may not have the positive knowledge required to send packets
of more than 576 bytes.

Second, if ARP fakery is used, the last-hop router may not be the last
hop, even if it thinks it is.  We may not want to cater to such
situations; I confess, though, that I've found ARP fakery exactly the
right technique to use when hanging a few SLIP lines off of a single
host -- otherwise, some IPs require me to use a separate network
number.

		--Steve Bellovin