--XAtoqsFAXbjZCl7Z
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Mar 08, 2006 at 10:33:06AM -0800, Sam Leffler wrote:
> You can turn off ACK's using WME but bidirectional traffic will fall=20
> down if the link is lossy as TCP doesn't handle this well (retransmit=20
> timers too coarse grained).

Even with SACK? (and, if you like, making assumptions of a continuous
TCP stream so that later in-flight segments can provoke SACKs, rather
than, say, interactive keystrokes which would have to time out).

At least to my naive initial guessing, it seems that freeing the
wireless layer of the responsibility of delivering packets in
sender-order could help throughput greatly if recovery was handled
well at a higher layer - and of course would have benefits for
datagram-oriented applications that are timing-sensitive but tolerant
of packet loss.

--
Dan.
--XAtoqsFAXbjZCl7Z
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (NetBSD)

iD8DBQFEDz9jEAVxvV4N66cRAlfEAKD4n2EH9Ig8dX+T3IKT7niNtvw7KwCg9UMp
Hht7BJ8mvezUCrNkfv+2f+0=
=V6mY
-----END PGP SIGNATURE-----

--XAtoqsFAXbjZCl7Z--