The MPLS WG Archive

[Ben Black <ben@layer8.net>]

On Thu, Dec 14, 2000 at 01:46:36PM -0500, Eric Gray wrote:
> Ben,
> 
>     The number of labels required to forward packets on
> a particular interface is a factor in determining the max
> MTU size for that interface.  If the "interface" is an LSP
> (or possibly stacked LSPs), then the LSR should take
> additional labels into account in determining the maximum
> MTU size.  Note that this does not mean that the LSR
> needs to know how many labels may be put on the packet
> as it is forwarded by other LSRs.  It only need to know
> the intrinsic maximum frame size for the LSP it will be using
> and the number of labels that it will put on.
> 
>     If there is a downstream LSR that adds additional labels
> but has a larger maximum frame size, then the frame size
> for the local LSR is unaffected by this.  It is generally best
> not to know more than you need to.
> 

This is exactly what I said.  The MTU determined should take into account
the additional labels.  I am not promoting a specific mechanism for
determining the MTU.  If the edge media MTUs are smaller than the core
media MTUs by enough to account for the labels, then the labels are
accounted for.  I am not clear on what you are arguing.


ben