Re: L2TPv3 MTU

From: Ronnie Angello <ronnie.angello_at_gmail.com>
Date: Sun, 28 Mar 2010 15:36:39 -0400

After further review, I only came up with 32 bytes for FR. I'm
assuming that they are referring to DLCI-to-DLCI mode, which only
supports 2 byte FR headers. Having said that, I'm assuming that this
is the breakdown -

20 bytes IPv4 delivery header
4 bytes L2TPv3 session ID
8 bytes L2TPv3 cookie
2 byte FR header (Q.922)

=34 bytes

On Sun, Mar 28, 2010 at 3:08 PM, Ronnie Angello
<ronnie.angello_at_gmail.com> wrote:
> Here is what I came up with after referencing Layer 2 VPN Architectures...
>
> FR over L2TPv3
> -----------------------
> 20 bytes IPv4 delivery header
> 4 bytes L2TPv3 session ID
> 4 bytes L2TPv3 cookie
> 4 bytes default L2 specific sublayer
>
> Ethernet over L2TPv3
> -------------------------------
> 20 bytes IPv4 delivery header
> 4 bytes L2TPv3 session ID
> 8 bytes L2TPv3 cookie
> 14 bytes Ethernet frame overhead
>
> Thanks,
> Ronnie
>
>
>
>
>
> On Tue, Mar 23, 2010 at 5:58 PM, sp-ie m <sp.ccie.me_at_gmail.com> wrote:
>> i'm trying to understand L2TPv3 header overhead from CCO -
>>
>> http://www.cisco.com/en/US/products/ps6587/products_white_paper09186a00800a8444.shtml
>> it says the following -
>>
>> The MTU in the IP backbone must be x bytes larger than the MTU that
>> operates on the pseudo-wire. The value for x is:
>>
>> 802.1Q = 50
>> Ethernet = 46
>> POS = 36
>> Frame Relay = 34
>> CHDLC = 36
>>
>> My understanding is for e.g. in Frame-relay , if i have default IP MTU
>> = 1500, and i generate an IP packet of 1500bytes,
>> 1. 1504-byte frame will be received on a Frame Relay attachment
>> circuit including the 4 byte Frame-Relay header
>> 2. 20 bytes L2TP IP packet header
>> 3. 12 bytes L2TPv3 session header
>> 4. 4 bytes - default Layer 2 specific sublayer (considering sequencing
>> is enabled)
>> Total = 1540 bytes.
>>
>> Can somebody explain Cisco's breakup? i cant understand how Cisco came
>> with th figure of "34 bytes" extra overhead over IP MTU.
>>
>> Similarly for Ethernet its listed as 46 bytes which is not clear.
>>
>> This comes from a real OEQ in the SP lab, hence i'm being lil extra meticulous
>>
>> Thanks.
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>> _______________________________________________________________________
>> Subscription information may be found at:
>> http://www.groupstudy.com/list/CCIELab.html
>>
>>
>>
>>
>>
>>
>>
>>
>
>
>
> --
> Ronald Angello
> Network Architect
> CCIE 17846
> CCDP, CCIP, CCNP
>

-- 
Ronald Angello
Network Architect
CCIE 17846
CCDP, CCIP, CCNP
Blogs and organic groups at http://www.ccie.net
Received on Sun Mar 28 2010 - 15:36:39 ART

This archive was generated by hypermail 2.2.0 : Thu Apr 01 2010 - 07:26:36 ART