Introduction to MPLS

Lessons Discussion
21

Thanks a lot for the great info

 

22

Good stuff, Rene’. I’m still working through deciphering the forwarding-table output, but we’ll see…

23

After using mpls instead of GRE do we still need IGP since mpls already use ip in the network layer. Is it possible to use mpls without igp.

Thank you for your support.

24

Hi Akram,

I think you refer to the PE-CE scenarios? In that case, you can’t:

  • We use MP-BGP to advertise VPN routes between the PE routers.
  • We establish the iBGP neighbor adjacency with loopback interfaces.
  • We use an IGP to advertise the loopback interfaces.

We still need an IGP there.

Rene

25

Hi Rene,

I am getting the same result but the output format was a bit different than yours in this lab, may I know what is the router platform and version you use?

Thanks
Elbert

26

Hi Elbert,

I used IOS routers in Cisco VIRL for this. These run IOS 15.x, it’s possible that routers running IOS 12.4 will have a different output for some show commands.

Rene

27

Hi Rene,

1st question, PE router is connected with a P router. We always talk about redundancy. Now, in this case, if P router is down then PE router can’t reach to it. How does SP solve this issue?

2nd question, do you have any plan for Nexus DC lessons?

Thanks and regards,

Nasim

28

Hi Rene,

These are very good but do you think you can create some videos as well and put it here so that its easier to listen while driving back home and understand the concepts. Thank you.

Regards,
Rankit

29

Hi Nasim,

In a real MPLS network, you will find redundant P routers in the core. In my examples, I try to use the minimum number of routers to explain something, to keep it simple. Nexus material is definitely something I am going to work on.

@Rankit I will add more videos yes.

Rene

30

hii Rene …your lessons are awesome…can you please add topics like
1)route-reflectors in MPLS network
2)hub and spoke topology in mpls
3)Ways of internet access in VRF’s
4)TE…etc
TIA

31

These are good topics indeed. I’ll add them in the future.

32

Hi, Rene

I had some issues with the first GRE tunnel configuration. I even inputted a static route on PE1 and 2 for traffic to use tunnel. Trace route would then work, but not ping. The MPLS config worked perfectly and you had great explanations above for it. Thank you!

33

Why does MPLS help so much with reliability and latency?

34

Hi Garrett,

Theoretically, label switching is faster than IP forwarding. About ~15 years ago that might have made a difference but nowadays, this doesn’t matter much.

When you use MPLS, you use a private network from the service provider which is far more reliable compared to a “regular” Internet broadband connection.

Internet connections don’t offer any QoS, it’s all best effort. MPLS does support QoS which ensures that delay sensitive traffic (like VoIP) is forwarded before other traffic.

Rene

35

Rene,

Could you provide the configuration file for the P routers in your architecture? I’m setting up lab so, I can get familiar with MPLS. I want to ensure I’m configuring the routing piece correctly.

Thanks
Tim

36

What is MPLS TE ?

Hi Rene, I want to learn MPLS TE now but I am not really sure of concept. Can you help please explain me the way you do other concept and if you can re-direct me to any LAB scenerio , I would practice my own on GNS.

thanks in adv

37

Hey Abhishek,
I would encourage you to check out our sister site, gns3vault.com. Over there you will find all kinds of cool labs written specifically for GNS3. While you can search for any lab related to MPLS, here is one specifically for MPLS TE

http://gns3vault.com/mpls/mpls-traffic-engineering-per-vrf/

38

HI Rene,

I am bit confused here, Why you have used IBGP between PE router as we are not using VPNV4 addresses, We are only using plain Ip customer addresses(No Vrf)
Incase of vrf we can use IBGP, pls explain.

39

19 posts were merged into an existing topic: Introduction to MPLS

40

Rohitendu,
In a provider network, BGP is required on the edges of the network (both to customers–to provide detailed Internet routing if requested by the custome, and to other providers). However, with a provider using a BGP-Free core, there is a danger of traffic getting black-holed. This means that a provider core will not have the full Internet routing table running in an IGP (because the IGP protocol can’t handle that many routes). So without an iBGP relationship between the PE’s the core would start dropping traffic (because the route would unknown).

iBGP relationships between all of provider’s exit points, helps each BGP router to determine where it needs to go to get to any particular destination on the Internet. The core’s job is to route this traffic to the chosen BGP exit point–not to know about the details of all Internet routes.

1 Like