Hi I just see that the mtu in lo is different from standard eth 1500 Any meaning of it? eg: Standard eth 1500 in linux. lo mtu 16436 in solaris. lo mtu is 8232 How about cisco / juniper loopback? Thank you so much
I think that the MTU on LO is pretty irrelevant in general. If it does matter, larger is probably better. Owen On Nov 17, 2010, at 9:08 AM, Deric Kwok wrote:
Hi
I just see that the mtu in lo is different from standard eth 1500
Any meaning of it?
eg: Standard eth 1500
in linux. lo mtu 16436
in solaris. lo mtu is 8232
How about cisco / juniper loopback?
Thank you so much
On 11/17/2010 11:08 AM, Deric Kwok wrote:
Hi
I just see that the mtu in lo is different from standard eth 1500
Any meaning of it?
You transfer huge amounts of data on loopbacks similar to sockets. Supporting large MTU's is appropriate, and given the virtual nature of loopbacks, is probably generally designed to handle the buffers that transfer the data.
How about cisco / juniper loopback?
Thank you so much
Juniper M120: Type: Loopback, MTU: Unlimited Cisco 7206 12.2SRE: MTU 1514 bytes, BW 8000000 Kbit/sec, DLY 5000 usec, Jack
Jack brings up a good point. MTU is basically pointless since packets never traverse any real interface....... So in theory the size can be anything...
Date: Wed, 17 Nov 2010 15:02:22 -0600 From: jbates@brightok.net To: deric.kwok2000@gmail.com Subject: Re: mtu question CC: nanog@nanog.org
On 11/17/2010 11:08 AM, Deric Kwok wrote:
Hi
I just see that the mtu in lo is different from standard eth 1500
Any meaning of it?
You transfer huge amounts of data on loopbacks similar to sockets. Supporting large MTU's is appropriate, and given the virtual nature of loopbacks, is probably generally designed to handle the buffers that transfer the data.
How about cisco / juniper loopback?
Thank you so much
Juniper M120: Type: Loopback, MTU: Unlimited
Cisco 7206 12.2SRE: MTU 1514 bytes, BW 8000000 Kbit/sec, DLY 5000 usec,
Jack
On Wed, 17 Nov 2010 16:23:54 -0500 Brandon Kim <brandon.kim@brandontek.com> wrote:
Jack brings up a good point. MTU is basically pointless since packets never traverse any real interface....... So in theory the size can be anything...
Not quite. You hit packet length field limits. IPv4 packets can't be larger than 65535, and IPv6 packets also can't be larger than 65 576 (40 byte IPv6 header + 2^16 payload), unless the jumbograms and the jumbo payload extension header is supported. Last time I checked, by setting the loopback MTU > 65 576, Linux, for example, doesn't support the jumbo payload extension header (or if it does, I didn't spend enough time finding out how to switch it on - a very large MTU didn't trigger it). That being said, with a 64K MTU on loopback, you can legitimately claim to get >10Gbps at home, as long as you don't mention how you're doing it ;-) Regards, Mark.
Thanks for the 411 Mark! Again, this NANOG list is such a valuable source of info and knowledge!
Date: Thu, 18 Nov 2010 08:18:10 +1030 From: nanog@85d5b20a518b8f6864949bd940457dc124746ddc.nosense.org To: brandon.kim@brandontek.com CC: jbates@brightok.net; deric.kwok2000@gmail.com; nanog@nanog.org Subject: Re: mtu question
On Wed, 17 Nov 2010 16:23:54 -0500 Brandon Kim <brandon.kim@brandontek.com> wrote:
Jack brings up a good point. MTU is basically pointless since packets never traverse any real interface....... So in theory the size can be anything...
Not quite. You hit packet length field limits. IPv4 packets can't be larger than 65535, and IPv6 packets also can't be larger than 65 576 (40 byte IPv6 header + 2^16 payload), unless the jumbograms and the jumbo payload extension header is supported. Last time I checked, by setting the loopback MTU > 65 576, Linux, for example, doesn't support the jumbo payload extension header (or if it does, I didn't spend enough time finding out how to switch it on - a very large MTU didn't trigger it).
That being said, with a 64K MTU on loopback, you can legitimately claim to get >10Gbps at home, as long as you don't mention how you're doing it ;-)
Regards, Mark.
I know in more recent Cisco IOS software (12.4.24T or later I think) the MTU of a GRE interface is that of the largest memory block on the box. I wouldn't be surprised if this is the same for loopbacks since they are a software concept. The logic behind this is that the largest frame the software interface can handle would be equal to the largest chunk of memory to put that frame in. On Wed, Nov 17, 2010 at 12:08 PM, Deric Kwok <deric.kwok2000@gmail.com> wrote:
Hi
I just see that the mtu in lo is different from standard eth 1500
Any meaning of it?
eg: Standard eth 1500
in linux. lo mtu 16436
in solaris. lo mtu is 8232
How about cisco / juniper loopback?
Thank you so much
participants (6)
-
Brandon Kim -
Deric Kwok -
Jack Bates -
Mark Smith -
Owen DeLong -
Pete Lumbis