06-16-2015 11:35 AM
Welcome to this Cisco Support Community Ask the Expert conversation. This is an opportunity to learn and ask questions about Configuring and Troubleshooting MPLS Traffic-engineering on Cisco IOS and IOS XR router to maximize network resource utilization with expert Vinit Jain.
IGP’s do not have traffic-engineering capabilities, Thus a major part of the network stays under-utilized. With MPLS Traffic-Engineering you can not only provision un-utilized links but can also overcome link or node failure conditions using Fast-Reroute. I will be helping you answer your queries on how to configure MPLS Traffic-Engineering and also help you troubleshoot the same.
Ask questions from June 22, 2015 through July 3, 2015
Vinit Jain, 4X CCIE #22854 is a Technical Lead in HTTS (High Touch Technical Support) team supporting customers in areas of routing, MPLS, TE, IPv6, multicast issues on IOS, IOS XE, IOS XR and NxOS code base. Vinit has been delivering trainings within Cisco on various technology as well as platform troubleshooting topics and is also a speaker at Cisco Live on Troubleshooting BGP. He has also written workbook on IOS XR fundamentals on Cisco Support Community. Vinit has expertise in troubleshooting Service Provider, Enterprise and Data Center environments and holds multiple certifications on programming and databases along with CCIE.
Vinit Jain was a recent speaker at Cisco Live in June 2015 on Troubleshooting BGP (BRKRST-3320). Click here for More Information
Find other https://supportforums.cisco.com/expert-corner/events.
**Ratings Encourage Participation! **
Please be sure to rate the Answers to Questions
Solved! Go to Solution.
06-22-2015 09:11 PM
Hello,
You are right with your understanding that RSVP is used for signaling and labelling for TE tunnels. You dont really require MPLS LDP to be enabled for TE Tunnels. LDP is enabled in the SP core for performing labelling and signaling operations for traffic that is not TE switched.
Also MPLS is required with TE when you have PE to P tunnels and your LSP gets broken in the SP core. In order to complete the LSP, we enable mpls on the Tunnel interface and on the tailend router, we configure "mpls ldp discovery targeted-hello accept".
Hope this answers your question.
Please let me know if you have any further questions.
Thanks,
Vinit
06-22-2015 08:37 AM
Good Day Vinit.
I'm thinking in begin to study MPLS TE, can you suggest me some books or links (videos, dorums) where I can review more information.
Tks.
Wilson
06-22-2015 08:47 AM
Hello Wilson
There is a great book on MPLS TE "Traffic Engineering with MPLS" by Eric Osborne and Ajay Simha. You can start with that. There is another book "Configuring MPLS on CIsco IOS Software" by Umesh Lakshman. It has good configuration examples too along with the explanation on control and data plane.
i have come across from interesting blogs which you can find below:
http://blog.codergenie.com/blog/search.aspx?q=mpls+trafffic-eng
there are multiple blog posts in the above link and i guess there are examples on IOS and IOS XR.
you can start by simulating few labs on GNS3 or VIRL.
Hope this helps.
Vinit
PS: Please rate the post if you find them useful.
06-22-2015 09:05 PM
Hello Vinit,
Could you let me know if MPLS is required to configure TE tunnels? From what i understand, RSVP is used as the signalling protocol for TE. Where exactly do we need MPLS with TE?
Thanks,
06-22-2015 09:11 PM
Hello,
You are right with your understanding that RSVP is used for signaling and labelling for TE tunnels. You dont really require MPLS LDP to be enabled for TE Tunnels. LDP is enabled in the SP core for performing labelling and signaling operations for traffic that is not TE switched.
Also MPLS is required with TE when you have PE to P tunnels and your LSP gets broken in the SP core. In order to complete the LSP, we enable mpls on the Tunnel interface and on the tailend router, we configure "mpls ldp discovery targeted-hello accept".
Hope this answers your question.
Please let me know if you have any further questions.
Thanks,
Vinit
06-25-2015 12:59 AM
Hello Vinit
Could you please help to explain is it possible provide few nodes protection by TE/FRR by cisco solution?
What is it best practices from cisco for provideing node protection FRR
I would be very appreciate for any links about FRR designing
thx
Gennadiy
06-25-2015 06:25 AM
Hello Gennadly
Yes, you can achieve multiple node protection using TE/FRR. You can define an explicit primary path and for the backup path, you need to configure an explicit path which are different than the nodes that you want to protect. Node protection is nothing but next-next-hop Tunnels and are configured in a similar way as Next-hop Tunnels with an exception that you manually define the complete backup path as well.
You can refer to the CCO documentation below:
http://www.cisco.com/c/en/us/td/docs/ios-xml/ios/mp_te_path_protect/configuration/xe-3s/mp-te-path-protect-xe-3s-book/mp-te-bfd-frr.html#GUID-964D987F-B555-44EB-9282-2073D33DE83C
http://www.cisco.com/c/en/us/td/docs/ios/12_0s/feature/guide/gslnh29.html
Hope this helps.
Vinit
PS: Please do rate the post if you find them useful.
06-25-2015 09:39 AM
We need to keep the below in mind when creating a backup Tunnel on the Head-End router for the Node protection.
"When using the exclude-address command to specify the path for a backup tunnel, you must exclude an interface address to avoid a link (for creating an NHOP backup tunnel), or a router-ID address to avoid a node (for creating an NNHOP backup tunnel)." - As per CCO.
Hope this helps.
Vinit
06-25-2015 11:54 PM
Vinit
Thank you so much for great assistance and explanation.
May i ask another question
LDP session protection and LDP synch with IGP - IS it to worth restricted to implement only one of them in production net ? If both need so why?
Sorry if i am fault
thx
Gennadiy
06-26-2015 08:23 AM
Hello Gennadly
Feel free to ask as many questions you have. I will be more than happy to answer them :) I would say the question on LDP sync and LDP session protection is a great question.
To answer this, below are my two cents:
++ Session protection keeps the Labels for the situation where the link comes back.
++ IGP sync prevents the change to a returning path as long as no labels are available
Usually, LDP IGP Sync alone is sufficient. LDP session protection helps by keeping LDP session up/running but igp-sync will still need link adj to come up/ready before declaring sync up. Given that there is *usually* no rush in putting a link back in routing (until it is really MPLS converged), customers usually delay igp-sync UP notifications by certain time (e.g. 45 sec). In such cases LDP session protection is not really needed.
P.S: The session protection feature will help avoiding any LDP control plane (label) convergence and may help scale system better if lots of label state needs to be exchanged among peers.
Hope this helps
Vinit
06-29-2015 10:04 AM
Vinit Thank you so much for so nice explanation
Another questions
1) RSVP using diffrent stiles i heard that cisco implemented only Fixed Flter ( i understood for point - to pont) and SE ( for multipoint- to -point) but what about WF ( wildcard for multipoint- to -point)? Implemented this style or no?
2) FEC for LDP based on destination prefix . i have read tha BGP assign only one label. but there are aritcles where write that BGP no agging label. where true ? is there any diffrence assigning process for i BGP vr EBGP and next-hop-self option in BGP setting?
3) QoS in MPLS onle EXP - that is small in order to make one - to -one mapping with dscp . Maybe there are some trick in configuration to make more soficticated transmission dscp from one site to another ( for example with assured classes)?
4) We are going to upgrade our networks that is based 7600 router and ES20 card we are looking for solution to connect diffrence Datacenters with Vmotion demands L2 connectivity - from you great experience and knowledge what technology is most perspective for these kind connectivity (EoMPLS VPLS EvPN OTV or something else . Our company is owner fiber communication and security features ( IPSEC or same ) is not important for us currently
Thank you advance for great discussion and i am sorry for a lot questions)
Gennady
06-29-2015 11:15 AM
Hello Gennady
Let me answer your questions one at a time:
regarding first question: we only support SE. rfc3209 only allows FF or SE, and we only do SE. See section 2.4. There may be something in future but i dont see any documentation / plans on the same.
I would like to understand what are we trying to achieve with WF style. We will still have to create Tunnels. So, is that really going to help?
06-30-2015 09:49 AM
Regarding Ques 2: i dont think there is any difference between label assigning process for iBGP or eBGP. i dont think Next-Hop-Self has anything to do with label assignment. BGP maintains its own table for the label it allocates. it can either allocate VPN label or can allocate label using send-label command.
06-30-2015 09:54 AM
Regarding Ques 3:
No matter if you you DSCP or IP precedence, only the first 3 bits of the IP header TOS field are used since the exp bits field is 3 bit long.
The mapping from DSCP to MPLS exp bits is:
0-7 to 0
8-15 to 1
16-23 to 2
24-31 to 3
32-39 to 4
40-47 to 5
48-55 to 6
56-63 to 7
I actually dont think you can have 1:1 mapping.
you can refer to the below CCO documentation for recommended way of doing the mapping:
http://www.cisco.com/en/US/tech/tk828/technologies_q_and_a_item09186a00800a43f5.shtml#wp27336
Hope this helps.
06-30-2015 09:56 AM
Regarding Ques 4: i think OTV and VxLAN BGP EVPN are a great option to choose. i recently presented on VxLAN BGP EVPN at Cisco Live. may be could help you with some decision making and configuration/troubleshooting stuff:
https://www.ciscolive.com/online/connect/sessionDetail.ww?SESSION_ID=83643
Hope this helps.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide