Community discussions

MUM Europe 2020
 
jonesy
newbie
Topic Author
Posts: 36
Joined: Tue Oct 07, 2008 1:31 am

What does /mpls ldp set distribute-for-default-route=yes do?

Wed Aug 19, 2009 4:37 am

Hi Guys,
I'm hoping someone will be able to help me understand what /mpls ldp set distribute-for-default-route=yes does.
It seems to add an entry in the MPLS forwarding table for the default route...

What I'm really interested in is which routers should have this option enabled, just the edges of the MPLS network (PE routers)? All the routers (P and PE)? just the P routers?
Thanks,
Andrew
 
Mplsguy
MikroTik Support
MikroTik Support
Posts: 226
Joined: Fri Jun 06, 2008 5:06 pm

Re: What does /mpls ldp set distribute-for-default-route=yes do?

Wed Aug 19, 2009 10:38 am

In general you are right - this setting controls if LDP distributes label binding for default route. Wether you should enable it, depends of how your routing is configured and what you use MPLS for. But you have to be careful with this setting: once label switching path is established across your network for some route, packet will travel across whole path - MPLS switching does not have "matching longest prefix" feature of IP routing. Consider setup:
R1-R2-R3-Internet
   |
   R4
R1 has only default route with gw R2, R2 has default route with gw R3 and more specific route to R4. All routers have "distribute-for-default-route" enabled.
If R1 wants to reach R4, it will make packet enter LSP established for default route (because it is the only route is has). R2 has default route to R3, therefore will MPLS switch this packet to R3 instead of R4. In case of IP routing, R2 would correctly route the packet.

In case you have complete routing table for your network on every router, this setting is safe to enable.

In case you use MPLS for something more that simple "speed up" of routing (e.g. IP VPNs, VPLS), enabling this setting can also cause issues if for some reason there is no binding for remote peer of particular service. Your VPN traffic can mistakenly get switched to wrong place (instead of remote peer, it will get switched across default route).
 
User avatar
ploquets
Member Candidate
Member Candidate
Posts: 128
Joined: Tue Nov 17, 2015 12:49 pm
Location: Uruguaiana, RS, Brazil
Contact:

Re: What does /mpls ldp set distribute-for-default-route=yes do?

Mon Dec 12, 2016 6:57 pm

In general you are right - this setting controls if LDP distributes label binding for default route. Wether you should enable it, depends of how your routing is configured and what you use MPLS for. But you have to be careful with this setting: once label switching path is established across your network for some route, packet will travel across whole path - MPLS switching does not have "matching longest prefix" feature of IP routing. Consider setup:
R1-R2-R3-Internet
   |
   R4
R1 has only default route with gw R2, R2 has default route with gw R3 and more specific route to R4. All routers have "distribute-for-default-route" enabled.
If R1 wants to reach R4, it will make packet enter LSP established for default route (because it is the only route is has). R2 has default route to R3, therefore will MPLS switch this packet to R3 instead of R4. In case of IP routing, R2 would correctly route the packet.

In case you have complete routing table for your network on every router, this setting is safe to enable.

In case you use MPLS for something more that simple "speed up" of routing (e.g. IP VPNs, VPLS), enabling this setting can also cause issues if for some reason there is no binding for remote peer of particular service. Your VPN traffic can mistakenly get switched to wrong place (instead of remote peer, it will get switched across default route).



Sorry about reviving this topic, but I'm curious about that as well and I could not understand what is this propose.

We have this scenario:
INTERNET (primary link)				INTERNET (secondary link)
		||										 ||
      EDGE 1  (CCR)						EDGE 2  (CHR)
	  	|							       	| 
  		|____________SWITCH ____________|
						 ||		|| 
						POP1     POP2    POP3 ..... etc etc etc                  
Every POP has its own vlan, on which both edges participate.
POPs are composed with 1 main router (RB2011, RB750... depends how many APs we have on at the POP) and all other APs connected to it.
Then we have a P2P bridge to connect this router to the Backbone, where is the switch and where it is connected to edges.

We use OSPF with Totaly NSSA area, every POP has its own area. Just to prevent route leak.
We don't want to have POP2 routes on POP1 devices, 'cause it doesn't make any sense, since traffic needs to reach edge to be routed back.

All those APs, we have filter rules to just receive a default route (from the NSSA...) and to redistribute connected routes, cause all APs are PPPoE Servers as well.

All Main POP routers, has filters to not receive a default route, and it has 2 default routes. First is CCR and second is CHR. With gateway monitoring.

We have enable MPLS on both EDGES, but, when I'm on a client from POP2 e.g. , packets take a lot to reach EDGE 2 like when trying to access winbox....
This only occur when MPLS is enabled on EDGE 2.

I don't know why this is happening. Its not MTU, I can guarantee. But, we had to disable MPLS on Edge 2.

Any advice about it? thanks!

Who is online

Users browsing this forum: No registered users and 16 guests