Re: Mods to Steve Deering's "DF" scheme

mogul (Jeffrey Mogul) Tue, 06 March 1990 18:38 UTC

Received: by acetes.pa.dec.com (5.54.5/4.7.34) id AA23944; Tue, 6 Mar 90 10:38:57 PST
From: mogul (Jeffrey Mogul)
Message-Id: <9003061838.AA23944@acetes.pa.dec.com>
Date: 6 Mar 1990 1038-PST (Tuesday)
To: fab%saturn.ACC.COM@salt.acc.com (Fred Bohle acc_gnsc)
Cc: mtudwg
Subject: Re: Mods to Steve Deering's "DF" scheme
In-Reply-To: fab%saturn.ACC.COM@salt.acc.com (Fred Bohle acc_gnsc) / Tue, 6 Mar 90 13:31:45 EST. <9003061831.AA15334@saturn.acc.com>

Fred writes:
    Jeff writes:
    >From: mogul@decwrl.dec.com (Jeffrey Mogul)
    >Message-Id: <9003052130.AA16912@acetes.pa.dec.com>
    >Date:  5 Mar 1990 1330-PST (Monday)
    >Subject: Modifications to Steve Deering's "DF" scheme
    >
    >I propose to fix this by using a "spare" IP Header bit (if the Proteon
    >gateways don't drop packets with unrecognized TOS bits, that's probably
						   ---
    Why not use the "RF" bit we were discussing in the FLAGS field?
    Right next to the "DF" bit.  That is the bit we were going to use
    before.  Besides, it applies to fragmentation, so let's keep it
    with the other fragmentation bits.

Because Proteon gateways are KNOWN to drop packets with that bit set.
A bug, but not one we want to tickle.  If they don't drop packets with
extra TOS bits set, that might be a safer approach.

Of course, we could also go ahead and use the bit in the FLAGS field,
and put Proteon on notice that their customers are going to be out
for blood.  But I'd rather let Proteon volunteer to accept this
solution than to cram it down their throat.

-Jeff