Re: draft-cameron-tmux-01.tx

["Donald E. Eastlake 3rd (Beast)" <[e-mail]>]

Dave,

Well, of course it is up to the sending host.  But if that host
thought it worth while to set the TOS on the original datagrams, I
would think the default would be that it would want the multiplexed
message to have only compatible datagrams and to have the same TOS as
they do.

I don't agree that TOS is useless.  Unused perhaps, but not useless.

Donald

From:  Dave Crocker <[e-mail]>
To:  "Beast (Donald E. Eastlake 3rd)" <[e-mail]>
Cc:  [e-mail]
Phone:  [phone]; fax: [phone]
In-Reply-To:  Your message of Tue, 02 Nov 93 02:56:17 -0500.
	               <[e-mail]> 
>well, I *thought* that I had read all the notes.
>
>More on TOS.
>
>You are, of course, correct that TOS is mandated.  However, it is a
>bloody useless field and there is no expectation that it will change
>anytime soon.
>
>Further, this is the originating host doing the ignoring.  It ought to
>know how important any of that stuff is.  To the extent the spec should
>comment on TOS (and I suspect that your raising the issue is a good
>indicator that the spec should say something specific) then it should
>caution the TMUX creator to synthesize and "appropriate" TOS value from
>those of the clients.
>
>Note that we've heard of TMuxing FTP and Telnet traffic.  They have
>different expected TOSs, yet the sender decides to multiplex them.  The
>sending TMux then gets to decide how to set the TOS value for the
>resulting single datagram.
>
>Dave