You can use either path vector or hop count independently. There are default values for each.

Joe

hello
I want know
(i) Is path vector & hop count are both needed simultaneously for loop detection in LDP
(ii) Is there any standard value of path vector limit & hop count value in LDP ?


Regards
Manoj Kumar Sharma

 Send free SMS to your Friends on Mobile from your Yahoo! Messenger. Download Now! http://messenger.yahoo.com/download.php

Shilpa,

For your first question an LER always does a lookup on the IP destination address of a packet regardless of whether or not the packet is meant to be encapsulated with MPLS or not. In other words, a router still functions like a router even with MPLS turned on. The destination of that packet will yield whether or not the packet will be sent out over a regular IP interface or encapsulated in MPLS and sent over an LSP.

(NOTE: In L2 VPN’s where you don’t do a L3 lookup this is different as usually a configured association is made between a CE-to-PE interface and an LSP)

For your second question, my statement above still applies. For example if your IP routing table yields the LSP as the next-hop to the loop back address of your IBGP peer then you send the BGP packets over the LSP.  For ISIS,OSPF, RSVP and LDP control packets those will still be sent out on the base POS/ATM/GE IP interface without MPLS encapsulation, but that is an exercise left up to the individual vendor. Note that if using hierarchical MPLS tunnels (LDP over RSVP) the LDP packets would ride over the MPLS LSP. Also, of note is that in all vendors you can control whether or not IP traffic and IGPs actually use LSPs for forwarding and next-hop computation with configuration knobs.

Hope this helps.

Thanks,

regards,
Shilpa

Shilpa,

Here is an example from an MPLS BGP VPN which shows the flow of information you are looking for. In this example the destination address would be 30.6.0.0. A route lookup in the virtual routing and forwarding instance table indicates that the next hop is an MPLS next-hop with a VPN label of 16 to be appended on to the packet, with a base label for the LSP of 19 which goes out over the POS0/0 physical interface. You are correct in that there are separate tables that maintain all of this information. The detailed routing table display ties it all together and shows the complete forwarding decision that is made for reaching this destination of 30.6.0.0

test:pe1#sh ip route vrf vpn1 30.6.0.0 detail

Protocol/Route type codes:

  I1- ISIS level 1, I2- ISIS level2,

  I- route type intra, IA- route type inter, E- route type external,

  i- metric type internal, e- metric type external,

  O- OSPF, E1- external type 1, E2- external type2,

  N1- NSSA external type1, N2- NSSA external type2

  L- MPLS label, V- VRF, *- via indirect next-hop

30.6.0.0/30 Type: Bgp Distance: 200 Metric: 0 Tag: 0 Class: 0

  MPLS next-hop: 132, label 16, VPN traffic, resolved by MPLS next-hop 60

    MPLS next-hop: 60, resolved by MPLS next-hop 123, peer 31.0.0.1

      MPLS next-hop: 123, label 19 on POS0/0 (ip19000002.mpls.ip [V:pe1]), nbr 11.13.0.1

This is an example of an MPLS forwarding table entry showing a label binding to the FEC 31.0.0.1. If you look at the display above you see that this label of 19 is displayed in the detailed “show ip route” output as the base label to append to packets headed to the next hop PE router 31.0.0.1 which originated the VPN route 30.6.0.0.

test:pe1#sh mpls ip binding 31.0.0.1

  31.0.0.1/32

    Out   19  neighbor 11.13.0.13

    Out   20  neighbor 11.14.0.14

Hope this helps in illustrating the forwarding decision.

Thanks,

Christopher Young
Resident Engineer
JNCIP-E ERX #9
(978) 973-0574
cyoung <at> juniper.net

---------- Forwarded message ----------
From: shilpa goel <shilpa07 <at> gmail.com>
Date: Jan 18, 2007 3:32 PM
Subject: Re: [MPLS-OPS]: basic doubt about LERs in MPLS
To: Christopher Young <cyoung <at> juniper.net>

Christopher,

Thanks for the reply. Can you give me a little more insight on the points given below?

You have mentioned that 'The destination of that packet will yield whether or not the packet will be sent out over a regular IP interface or encapsulated in MPLS and sent over an LSP'
As I understand routers maintain 2 routes - one in IP forwarding table for some destination say 192.168.1.0 and one in MPLS LFIB with the same destination address. I am considering the case of LDP protocol setting up LSPs in response to route updates by IGPs. Now when a LER interface receives an IP packet with destination 192.168.1.0, will it consult some common table which will indicate which of the 2 tables mentioned above (IP or MPLS) will be used for routing of the packet? If so, how is this common table configured?

Also is it "possible" or "advisable" for all packets (control/protocol or management) generated at the router to be IP routed by default? What is the usually adopted implementation?

thanks,
Shilpa