VXLAN BGP EVPN- L3VNI (Episode 2)

In this video, I dig deeper into "Control plane Learning" approach for VXLAN using BGP EVPN. I focus on "Routing" i.e INTER-VNI communication with a modern Leaf-Spine topology on EVE-NG

[@sachinporwal5622]

This is a unique and awesome explanation on such complex topic.......simply great...

[@bineeshphilip278]

Rally great explanation, one of the best video 👍👍

[@briandsouza1550]

Every second was informative!! Thank you

[@kerryk232323]

excellent video, this video helped me prepare for an interview and they asked me if I learned this at Cisco when I worked there

[@evanhairston5872]

Excellent video, thank you!

[@bayramagc6592]

Thank you so much for these great videos.

[@UncoveringTheHiddenTruth]

a very difficult topic explained in simple terms. Great work done

[@easytechnology-FK]

Great explanation, really helpful and appriciate

[@sukumarmondal2216]

This is awesome informative video

[@maltijadhav1689]

Thanks a lot for such a informative video

[@snehasharath5085]

Excellent explanation! Specially regarding Asymmetric and Symmetric IRB. I had read the Cisco design guide on this but could not understand it completely. Your content helps a lot! Thank you.

[@BitsPlease]

Thank you, glad you found it useful.

[@kimsvalkvist3161]

Great explanation of how it all works! helped me alot to understand how it all comes together! Tnx for that! :D

[@shahadatanwar4109]

Great explanation

[@rajbas1628]

Nice work !!!

[@JoyEnrich]

The Best ,Thank you so much :)

[@venuonduty5778]

Great, I like this video

[@parvez7077]

one of the best video on vxlan implementation with concept..crystal clear explanation

[@BitsPlease]

Thanks Munavvar.

[@parvez7077]

@@BitsPlease one question - under evpn section ,we generally mention the L2VNI but I did not see in your video. Is that for L2 route exchange? ex: [evpn vni 100800 l2 rd auto route-target import auto route-target export auto]

[@BitsPlease]

@@parvez7077 The L2VNI section is covered in the previous video. You can find it in the vxlan playlist on the channel. Further my GitHub repo has all the configs in one single place

[@Mallunetworking]

really appreciate

[@mubashir1976]

great!!!!

[@bestsaurabh]

Doubt: In the case of symmetric IRB, how is L3VRF populated? Does it have a mapping of destination IP/subnet and destination VTEP? How is routing working here, based on what I will forward the packet to destination VTEP? Also do we map all the L2 VNIs present on a vtep to a common L3 VNI. Can you please elaborate on that? Thanks for the awesome explanation :)

[@123prova]

This video has a lot of useful information. The reason why we need a new VRF is clear but the creation of a new VNI and a new vlan 999 + SVI without an IP address is a bit weird. Vlan 10 is also belonging to two different contexts (the new VRF and the previous VNI) makes it a twisted concept. Perhaps a bit more detailed diagram of 16:52 would have helped there to understand the internal connectivity . This video is much more than a starting point and very useful to grasp a good amount of concepts. Thanks .

[@vasantnagarajan3463]

Great Video!! Had a query with the bgp config, why aren't we mentioning/using the "neighbor commands" inside the VRF while configuring the network?

[@123prova]

if I remember correctly that was already configured in the first video (59:38)

[@pixelmage3523]

Hey Amazing explanation again as usual ,I just have one question,You advertized 10.10.1.0 subnet under vrf TENANT in BGP,What purpose is this serving ? Wont the VTEP s learn these IP addresses anyway from remote VTEP s or from hosts attached to them ?I just can't get my mind around it . Please help .Is it because when a VTEP learns a Mac address from host,It's just put in base address family (With no vrf) and we need routes to hosts inside the VRF routing table for stuff to work ? ??

[@vasantnagarajan3463]

Excellent video, just a query regarding the vrf config, i see that you have not put any neighbor command in the vrf like we do in IOS, Is this different for NXOS? Does the vrf take the global neighbor command for sending the RT's?

[@BitsPlease]

The BGP config is pretty similar to MP-BGP used in MPLS L3VPN - kzbin.info/www/bejne/maKmZJidgdZ9bs0. Neighbor inside VRF comes into play while doing VRF-LITE - kzbin.info/www/bejne/mqi7g2drnrSZd9k. These 2 videos on my channel should help you see the differences.

[@Bormanb23]

Thank you! what if I had hosts on non VXLAN networks and wanted to "talk" to hosts in VXLAN networks? the VNIs are on a separate VRF (L3 VNI VRF)

[@BitsPlease]

I do talk about that in the next video in the series where we connect vxlan to an external network

[@lethedan]

L3VNI Routing between VNIs happen at the Spine nodes - correct? Thanks

[@BitsPlease]

Nope at the leaf. That's why we called it IRB - integrated routing and bridging

[@anshulverma8023]

Hi BitsPlease, Please help me in understand how the VTEPs learn about each other, I mean lets assume I connected the fabric (Spines/Leafs) for the first time and configured the underlay network, Now as of now there are no hosts trying to communicate to each other. I understand that I need to configured bgp-evpn or use flood & learn for host -Vtep learning, I guess I am confused between host learning and VTEPs learning. Please help

[@BitsPlease]

Anshul Verma In Flood & learn , the VTEP discovery happens via data plane. The ARP is flooded to the multicast group and the remote VTEP will learn about the source VTEP via the flooded traffic. In BGP EVPN, the VTEPs will know about each other beforehand since we put that in the configuration

[@mubashir1976]

Assuming we are using BGP EVPN for control plane, VTEPS are discovered as a by product of BGP routes update exchanged. Important: No data plane traffic has to be present for some BGP updates to trigger, for example: route tye 3, route type 4 are generated without data plane traffic. In static VXLAN where we do not use any control plane to learn MAC/IP routes. we configure statically VTEPS IPS on each VTEP. Hope its clears up.

[@mubashir1976]

At 9:41, you mentioned following: When Irb associated with v5000, receives a packet destined to 20.1.1.1, it will look for MAC VRF associated with V2000. How does router determine which MAC vrf to look? Let say we have 4000 MAC VRFS, then 20.1.1.1 is associated with MAC VRF 20, how will router know to look ino MAC VRF 20 out of 4000 MAC VRFS? Thanks!!

[@BitsPlease]

It's a typical L3 routing lookup. In Asymmetric IRB, the router tries to look up 20.1.1.1 and finds it's with associated vlan 20 which inturn is part of MAC VRF 6000. In Symmetric IRB, vlan 20 would not exist hence it is routed off the L3 VNI (IP VRF)

[@mubashir1976]

@@BitsPlease Crystal clear!!

[@bineeshphilip278]

Why are we advertising network under Tenant1 vrf ? There is no bgp neighbor under Tenant1 vrf? How does this route will be advertised?

[@BitsPlease]

Can you provide the timestamp of the video you are referring to

[@bineeshphilip278]

It is 35:31. And how can we advertise l2vpn evpn under ipv4 address family? Not getting what are we trying to achieve in this step? Could you pls explain?

[@BitsPlease]

@@bineeshphilip278 This video is a continuation of the previous one in the series. The rest of the BGP config is already on the box as it was done in the Episode 1 of the series - kzbin.info/www/bejne/nJK4lWOsmLetf5o