The MPLS WG Archive

["Lipovetsky, Sergey" <slipovetsky@virata.com>]

Hi David,

Thank you for your response.
I will try to restate the things. Please correct me if something is wrong.

Suppose we have two ingress routers (IP addresses: ingr1, ingr2) and two
egress routers (IP addresses: egr1, egr2)
and several TE tunnels defined by the following session objects:

     extended tunnel ID        tunnel ID     IPv4 tunnel end point address
1)        ingr1                           1                       egr1
2)        ingr1                           1                       egr2

3)        ingr2                           1                       egr1   
4)        ingr2                           1                       egr2

5)           0                             1                       egr1
6)           0                             1                       egr2

All these TE tunnels can co-exist without any collisions.
In cases 1-4 tunnel ID is allocated for a given pair of ingress-egress
nodes.
In cases 5-6 tunnel ID is allocated for a given egress node.

Sergey
 
> -----Original Message-----
> From: David Charlap [mailto:david.charlap@marconi.com]
> Sent: Wednesday, August 01, 2001 4:19 PM
> To: 'mpls@uu.net'
> Subject: Re: Doubts about MPLS-TE Mib
> 
> 
> "Lipovetsky, Sergey" wrote:
> > 
> > How can we say that tunnel ID is locally unique to the ingress
> > node in this case?
> 
> Because RSVP has no mechanism for synchronizing tunnel IDs across the
> network.  Therefore the value is, by definition, only locally unique. 
> This should be obvious.
> 
> > This is true only when the extended tunnel ID is not zero.
> 
> The uniqueness of tunnel ID is not affected by the value of 
> other fields
> in the SESSION object.
> 
> You seem to be confusing the tunnel ID with the entirity of 
> the SESSION
> object.
> 
> > If the extended tunnel ID is zero, some global synchronization is
> > required.
> 
> It would be a good idea, but the protocol does not provide for any
> mechanism.  And RSVP will work just fine with or without said
> synchronization.  Your requirement is nothing more than wishful
> thinking.
> 
> > Can we say that tunnel ID is local unique to the egress node in this
> > case?
> 
> I don't see what your preoccupation with this is.  You seem die-hard
> determined to change the RSVP spec such that it fits the MPLS-TE MIB.
> 
> > This is different from CR-LDP, where lspid is really locally unique
> > to the ingress node and does not require any synchronization.
> 
> CR-LDP doesn't allow two LSPs to share resources either, so the entire
> argument is moot there.
> 
> -- David
>