The MPLS WG Archive

["Dutta, Pranjal" <pdutta@riverstonenet.com>]

Earlier mail got bounced..resending again…

 

Hi Eric,

                 This answers well to the concern. It would depend whether we would like to have an all MPLS network spanning via LDP or not.

It is well said that aggregation can be turned off or on based on the whether we would like to have an inter-area LSP or not. At this point I would

like to go by that at Rd, LDP is unaware of whether the prefix entries found in its route table for downstream has been summarized to Ru. As David said

it has a side effect with LSP-Ping.

 

Regards,

Pranjal

 

---

From: Eric W Gray [mailto:ewgray2k@netscape.net]
Sent: Thursday, 13 April, 2006 9:13 PM
To: Raveendra Torvi
Cc: DECRAENE Bruno RD-CORE-ISS; mpls@ietf.org
Subject: Re: [mpls] One LDP Implementation specific question of receivelabelmapping for prefix FECs

 

Let me answer just the first part of this and let the rest take care of itself... :-)

LSPs are [terminated] at aggregation boundaries.  This may not be desirable in case of
an all-MPLS network.

To some extent, every box needs to "make up its mind" whether it is a router or a switch.
It is important in interoperating between boxes that they are each self-consistent.

The aggregator router could decide not to aggregate routing information.  That way, it can
distribute more specific labels and not worry about inconsistencies between the routes it
has advertised and the labels it distributes.

However, there are reasons why routers aggregate routing information, and I think it may
be difficult to justify why those reasons would not apply equally well to label distribution.

It may very well be the case that routing aggregation is not completely justified in specific
deployment scenarios. In that case, perhaps it would simply be better to "turn off" route
aggregation.

--
Eric

 

_______________________________________________
mpls mailing list
mpls@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/mpls