Route Leaking And LDP Extension for inter-area lsp (RFC 5283)

Service Providor networks usually will used   link-state  ISIS or OSPF as IGP with hierarchical areas and route summarization (dependent of their size)  .
networks based  ip/mpls  with classic LDP signaling  summary route and the areas broken the Lsps connectivity.
this design provides challenges for  provision  of services end to end.

There are many   techniques to  overcome this situation they can call "Seamless Mpls" one of them  is the RFC 5283 to solve the summarization.

Normally for LDP binding LSR needs to ensure there is an exact match for X in the RIB anything else is dropped,the RFC suggest changing this verification procedure to the longest match if there prefix Y/24 in the RIB and the X/32 is one of the subnet the X/32 keep lable mapping and propagated in LFIB (you need that all LSR in lsp  will enable RFC ),so you keep the  end-to-end connectivity

on the other hand the   RIB and the LIB prefixes are still summarized

the drawbacks of this approach is that the LFIB size growth, now it depend what your prefer the maintaining LFIB is less compared to maintaining IGP databases :-)

    :For example

 R-Pe2 try to built lsp to PE2 but the LDP did not have the prefix 172.20.20.1 in the RIB cause in the area of level-1  have only default route from PE1 in result the ldp did not mapping this prefix .In the other way PE2 try to built lsp to the R-pe2 but in the RIB  there are prefix 10.11.11.0/24 and not the exact prefix 10.11.11.2/32 ,So the ldp did mapping this prefix as well

The solution:


PE1 configure with specific leaking level 2 to level 1 of the loopback of PE2
In result R-pe2 have in the RIB and in the LIB/LFIB the prefix 172.20.20.1/32 and the LSP can be  built.
In the other way the PE1/P/PE2 configure with the support of RFC 5283 so in the LFIB of PE2
The FEC are built in result the lsp built .

Next post I will write about other techniques  

Further Reading:

LDP Extension for Inter-Area Label Switched Paths



MPLS LDP-IGP Synchronization

Problem:

ldp establishes mpls lsps path that determined by the IGP best route ,if the lsp not formed between the PE routers, a blackhole exists and services that depending on mpls forwading will fail,this feature will help to avoid packet lost in this situations:

1.when the IGP adjacency is up and the LDP sessions are not establishes or the lable-bindings not excheanged with the neighbor.

2.when someone remove the ldp under the interface and the LDP session not form (Configuration error)

 

Solution:

the solution is to discourage a link from being used as long as the LDP sessions are not established.

so when you enable the feature of IGP-LDP sync and the LDP is not fully operational on a given link,

the IGP will advertise the cost of the link to be maximum in order to avoid any traffic to flow through it,

The actual value advertised in OSPF (65535).

The actual value advertised in IS-IS regular metric (63) and in IS-IS wide-metric is (16777214).

 in Cisco :

enter the command  mpls ldp igp sync  under the IGP (ospf or isis) the IGP waits indefinitely for LDP session to be fully operational ,To limit the length of time the IGP session must wait, enter the no mpls ldp igp sync holddown command

in ALU :

enter the command ldp-sync-timer seconds  .

 

source:

http://www.cisco.com/en/US/docs/ios/12_0s/feature/guide/fsldpsyn.html#wp1100282

http://infodoc.alcatel-lucent.com/html/0_add-h-f/93-0073-10-01/7750_SR_OS_Router_Configuration_Guide/router-CLI.html

MPLS Label Distribution Protocol Fast Re-route (LDP FRR) using ISIS as IGP.

Introduction:

LDP is widely deployed to signal LSP due to its simple operational model.
Since Ldp establishes LSPs along IGP routed paths,its failure recovery is gated by the IGP re-convergence.this feathers perforn FRR of LDP LSPs that reatains the simple operational model of LDP.
The goal is to provide 100% coverage for all failure with recovery charateristics (link /node failures) similar to the methods in Reservation Protocol - Traffic Engineering FRR.

When FRR is configured on a node, the node pre-computes primary NHLFEs for all FECs and in addition it will pre-compute backup NHLFEs for all FECs. The backup NHLFE corresponds to the label received for the same FEC from a Loop-Free Alternate (LFA) next-hop (see also RFC5286). Both primary NHLFEs and backup NHLFEs in ALU 7X50 are programmed in the IOM/IMM which makes it possible to converge very quickly.

ISIS LFA:

Implementation :

 

The 7x50 software has implemented inequality 1 (link criterion) and inequality 3 (node criterion) of RFC5286. Similar to the Shortest Path Tree (SPT) computation that is part of standard link-state routing functionality, also the LFA next-hop computation is based on the IGP metric.

The underlying LFA formulas look like this :

Inequality 1 : [SP(backup NHR, D) < {SP(backup NHR, S) + SP(S, D)}]

Inequality 3 : [SP(backup NHR, D) < {SP(backup NHR, PN) + SP(PN, D)}]

With ‘SP’ = ‘shortest IGP metric path’, ‘NHR’ = ‘next-hop router’,’D’ = ‘destination’, ‘S’ = ‘source node or upstream node doing the actual LFA next-hop computation’ and ‘PN’ = ‘protected node’. Inequality 3 rule is stricter than inequality 1 rule. if the formulas does not exist the prefix did not protected.

Configuration :

The IGP CLI configuration will look like this :

*A:7450-lab3# configure router isis loopfree-alternate

After enabling LFA inside the IGP context, FRR needs to be enabled within the LDP context :

*A:7450-lab3# configure router ldp fast-reroute

Several show commands are possible to display LFA information:

 *  ‘show router isis statistics’ gives you an idea on the number of LFA runs on a specific node.

 *‘show router isis lfa-coverage’ will perform a mathematical calculation  between the number of nodes and    IPv4/IPv6 routes in the network versus  present LFA next-hop protections:

*A:7450-lab3# show router isis lfa-coverage

===============================================================================

LFA Coverage

===============================================================================

Topology Level Node IPv4 IPv6

-------------------------------------------------------------------------------

IPV4 Unicast L1 0/0(0%) 9/9(100%) 0/0(0%)

IPV4 Unicast L2 4/4(100%) 9/9(100%) 0/0(0%)

===============================================================================

*   ‘show router isis spf lfa detail’ gives you a reference to LFA protection type (link or node).

* ' show router isis routes alternative'  give you the next hop backup

*  'show router ldp bindings active prefix 172.30.1.1/32' we can see the backup lable for the fec