Re: [mpls-linux-general] LSP binding, nesting automatically

[James R. L. <jl...@mi...>]

This a fairly vague question. Each protocol/MPLS service has it own
mechanisms for doing them.  Pick one particular one and I can explain it
to you.

Here is an example to refer to:

Layer 2 cross connects across an LDP created LSP that utilizes a RSVP
traffic engineered core.

In this case you will have:

RSVP-TE creates 'tunnels' (think of a big pipe) that control how large
amounts of traffic move from P to P router.  LDP then runs in direct mode
from the PE to the P routers, and indirect mode across the RSVP-TE 'tunnels'
(from P to P routers).  The resulting LDP LSPs run from PE to PE but are
tunneled inside of the 'big pipe' to cross the core.  Ontop of the LDP LSPs
another indirect LDP session is created to exchange labels for the layer 2
cross connect. (think of these as 'application labels')

Which part of this (or another example) do you want me to explain?

On Fri, Apr 18, 2003 at 09:41:55PM +0800, Sun Jianwei (Dr) wrote:
> hello all, 
>     when a lsp is established, how can it broadcast to network, so that the lsp can bind with other FECs or  nest other lsps?  I think one aproach is to register the lsp as a virtual network interface, unnumbered link, then it can be broadcasted through routing protocols. is there any implementation of such mechanism? 
> 
>     mplsadm provides tunnelling(or lsp nesting) mechanism, but it is manually configured. how to make it to be dynamically? is there any related IETF RFC or draft?
> 
> thanks.
> Jianwei   
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu