| Author |
Message |
Swingman
Guest
|
 Posted:
Sun Jan 02, 2005 5:25 am Post subject:
PPPoE and "allowable MTU" |
|
|
Just noticed that PPPoE is limited to 1492 in my Speedstream
5100b modem (SBC Yahoo DSL service). My router and OS are
set for 1500 and I wonder if they should be changed to 1492
(or some other number) to optimize the connection? |
|
| Back to top |
|
 |
Neil W Rickert
Guest
|
| Posted:
Sun Jan 02, 2005 5:25 am Post subject:
Re: PPPoE and "allowable MTU" |
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
"Swingman" <sbt@silcom.com> writes:
| Quote: | Just noticed that PPPoE is limited to 1492 in my Speedstream
5100b modem (SBC Yahoo DSL service). My router and OS are
set for 1500 and I wonder if they should be changed to 1492
(or some other number) to optimize the connection?
|
Check whether you are having problems.
If your OS does path MTU discovery, then it shouldn't matter.
If your OS sends 1500 byte packets, and those are fragmented by the
router or modem, then that might slighly lower your upstream speed,
and setting to 1492 would correct that. However, it it uses path mtu
discovery, then your OS will have discovered that 1492 is the max, so
it will only send 1492 byte packets.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.3.91 (SunOS)
iD8DBQFB10/xvmGe70vHPUMRAs0VAJ9+gofhD2V6Ne/3+x54PhE3+NNbeQCg7Fis
yB9eQHLw1Oxips4qgbkZY7A=
=p4+z
-----END PGP SIGNATURE----- |
|
| Back to top |
|
|
David H. Lipman
Guest
|
| Posted:
Sun Jan 02, 2005 5:43 am Post subject:
Re: PPPoE and "allowable MTU" |
|
|
No. The WAN side should be 1492 because there is an 8byte overhead in tunneling IP with
PPPoE. MTU should remain 1500 on the LAN side as this is the "standard" for Ethernet.
However, it has been noted in the past that "IF" you have problems retrieving/sending email
or posting to UseNet you *may* have to set the LAN nodes MTU setting to 1492. However, it
should only be done if there is a problem.
--
Dave
"Swingman" <sbt@silcom.com> wrote in message news:10teft1drjd4t69@news.supernews.com...
| Just noticed that PPPoE is limited to 1492 in my Speedstream
| 5100b modem (SBC Yahoo DSL service). My router and OS are
| set for 1500 and I wonder if they should be changed to 1492
| (or some other number) to optimize the connection?
|
| |
|
| Back to top |
|
|
Swingman
Guest
|
| Posted:
Sun Jan 02, 2005 9:26 pm Post subject:
Re: PPPoE and "allowable MTU" |
|
|
"Neil W Rickert" <rickert+nn@cs.niu.edu> wrote in message
news:cr7j5l$flb$2@usenet.cso.niu.edu...
| Quote: | -----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
"Swingman" <sbt@silcom.com> writes:
Just noticed that PPPoE is limited to 1492 in my
Speedstream
5100b modem (SBC Yahoo DSL service). My router and OS are
set for 1500 and I wonder if they should be changed to
1492
(or some other number) to optimize the connection?
Check whether you are having problems.
If your OS does path MTU discovery, then it shouldn't
matter.
If your OS sends 1500 byte packets, and those are
fragmented by the
router or modem, then that might slighly lower your
upstream speed,
and setting to 1492 would correct that. However, it it
uses path mtu
discovery, then your OS will have discovered that 1492 is
the max, so
it will only send 1492 byte packets.
|
Thanks for both replies. I'm getting good throughput, and I
use WindowsXP with Path MTU Discovery ON so I think all is
well. |
|
| Back to top |
|
|
T. Sean Weintz
Guest
|
| Posted:
Wed Jan 05, 2005 3:00 am Post subject:
Re: PPPoE and "allowable MTU" |
|
|
Swingman wrote:
| Quote: | Just noticed that PPPoE is limited to 1492 in my Speedstream
5100b modem (SBC Yahoo DSL service). My router and OS are
set for 1500 and I wonder if they should be changed to 1492
(or some other number) to optimize the connection?
|
Yes. Otherwise you (in some circumstances) cause the speedsteam to have
to fragment the packets. This would only happen with packets between the
size of 1492 and 1500.
T. Sean Weintz |
|
| Back to top |
|
|
Robert Redelmeier
Guest
|
| Posted:
Wed Jan 05, 2005 5:36 am Post subject:
Re: PPPoE and "allowable MTU" |
|
|
T. Sean Weintz <strap@hanh-ct.org> wrote:
| Quote: | Yes. Otherwise you (in some circumstances) cause the
speedsteam to have to fragment the packets. This would only
happen with packets between the size of 1492 and 1500.
|
I tested my setup (SS 5360, Siemens 2602 router, Linux)
and found a small (4%) advantage for 1492 (data below).
Since the fragmentation happens on the fast 10baseT network
between my router & modem (or perhaps within the modem itself),
it doesn't make much difference.
-- Robert
$ for i in `seq 1440 8 1500` ; do ping -qc 10 -s $i gw2 ; done
PING gw2 (70.240.223.254) 1440(1468) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9011ms
rtt min/avg/max/mdev = 56.755/57.230/57.635/0.261 ms
PING gw2 (70.240.223.254) 1448(1476) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9011ms
rtt min/avg/max/mdev = 56.623/57.146/57.469/0.416 ms
PING gw2 (70.240.223.254) 1456(1484) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9011ms
rtt min/avg/max/mdev = 56.770/57.284/58.066/0.416 ms
PING gw2 (70.240.223.254) 1464(1492) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9011ms
rtt min/avg/max/mdev = 56.900/57.380/57.788/0.440 ms
PING gw2 (70.240.223.254) 1472(1500) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9012ms
rtt min/avg/max/mdev = 59.047/59.766/60.740/0.496 ms
PING gw2 (70.240.223.254) 1480(1508) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9013ms
rtt min/avg/max/mdev = 59.881/60.153/60.432/0.318 ms
PING gw2 (70.240.223.254) 1488(1516) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9002ms
rtt min/avg/max/mdev = 59.388/59.984/60.494/0.452 ms
PING gw2 (70.240.223.254) 1496(1524) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9010ms
rtt min/avg/max/mdev = 59.273/59.911/60.636/0.423 ms
and normally:
$ ping -qc 10 gw2
PING gw2 (70.240.223.254) 56(84) bytes of data.
10 packets transmitted, 10 received, 0% packet loss, time 9011ms
rtt min/avg/max/mdev = 10.702/11.322/12.552/0.555 ms |
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in