Point-to-multipoint vs point-to-multipoint non-broadcast

Networking CCNP Enterprise
1

Hi
what is the difference between
point-to-multipoint vs point-to-multipoint non-broadcast ?

i dont understand what is point-to-multipoint non-broadcast used for ? i know its cisco proprietry network type which means it`s deployed on cisco devices only , but is that means if i want to deploy this type on one side the other side must be supported point-to-multipoint non-broadcast also or the network type will be mistmatch between both sides ? which means both sides must be cisco device in order to support the same type point-to-multipoint non-broadcast ?

2

Hi major133,

So the “network type” in OSPF controls a few things, such as:

  • How many adjacencies can occur on the link?
  • Are updates sent as multicast or unicast?
  • Is there a DR/BDR election?
  • What is the HelloInterval and DeadInterval?

In the case of “point-to-multipoint” network type, it has the following attributes

  • Multiple adjacencies are supported on the link
  • Updates are sent as Multicast to 224.0.0.5
  • There is no DR/BDR election
  • HelloInterval is 30 sec, DeadInterval is 120 sec

In the case of “point-to-multipoint non-broadcast” network type, it has the following attributes

  • Multiple adjacencies are supported on the link
  • Updates are sent as Unicast
  • There is no DR/BDR election
  • HelloInterval is 30 sec, DeadInterval is 120 sec

The design idea of P2M Non-Broadcast is that you may have a circuit, for example like a VPLS, that is emulating a LAN over the WAN. From the CPE perspective, it looks like a flat LAN where everyone is in the same broadcast domain, and hence the same IP subnet. On the WAN side however, it’s really a collection of Virtual Circuits, usually running over MPLS. In order to add Multicast support to an emulated LAN, the PE’s would have to fully mesh their Virtual Circuits, or use some sort of hierarchy of reflection such as an H-VPLS.

In early days of VPLS hardware, there was a strict small limitation on how many VCs could be supported, due to linecard limitations. Therefore not all VPLS circuits would inherently support IP Multicast transport. Multicast would be available on the circuit for an additional circuit charge ($$$) if you needed it.

Enter workarounds like OSPF P2M Non-Broadcast. We run OSPF as NOT an emulated LAN (which would be OSPF network “broadcast”), but all the Hellos and Updates are sent as IP Unicast, based on you manually specifying the “neighbor” statement under the OSPF process.

The disadvantage of this on the CPE side is that you’d have to update each router’s config with an extra “neighbor” statement everytime you want to add a new site to the VPLS, but the advantage is that you don’t need require IP Multicast transport.

Make sense?

Thanks!

Brian
bmcgahan@ine.com

1 Like
3

Hello great brian, thanks for your reply
could you please give me another examples because i dont know what is VPLS?could you explain using frame-relay example,i know what is the frame relay but i dont know what is vpls