Re: IP: optimize as a router not host

Matthias Urlichs (smurf@smurf.noris.de)
Mon, 18 Mar 1996 21:07:57 +0100

In linux.dev.kernel, article <314D35B9.698D@et.tudelft.nl>,
Rogier Wolff <r.e.wolff@et.tudelft.nl> writes:
>=20
> Suppose a sender sends packets 1,2,3,4 and 5. At the recieving end
> you get 1,2,4,5. What you do is you ack that you got packet 1 and two=
.=20
> When you get 4 and 5 you ack that you got 2 again and again. When a=20
> sender gets these it is supposed to conclude that packet 3 got lost
> and try resending that.=20
>=20
There is a RFC which implements selective reject -- i.e., in the ack, y=
ou
add some TCP options which tell the sender that you do have packets 4 a=
nd 5
(or rather, the sequence numbers for which 4 and 5 contained data).

> In short only the endpoint can determine the correct action to take
> when a bad-checksum-packet arrives....
>=20
> (maintaining a table of active connections through the router and
> recording the state of all those is unfeasable: we have a linux-route=
r
> here that has thousands of active connections going through it.)
>=20
We have the IP masquerading stuff which does exactly that.

In the general case, i.e. if you don't need to do masquerading, of cour=
se,
I agree with you -- the added overhead certainly isn't worth it.

--=20
Ambition: The glorious frailty of the noble mind.
-- Hoole
--=20
Matthias Urlichs \ XLink-POP N=FCrnberg | EMail: urlichs@smurf.=
noris.de
Schleiermacherstra=DFe 12 \ Unix+Linux+Mac | Phone: ...please use =
email.
90491 N=FCrnberg (Germany) \ Consulting+Networking+Programming+etc'i=
ng 42
PGP: 1B 89 E2 1C 43 EA 80 44 15 D2 29 CF C6 C7 E0 DE=20
Click <A HREF=3D"http://smurf.noris.de/~smurf/finger">here</A>.