Hi Cem,
Glad to hear you like it!
Frame-relay is an old WAN protocol. It’s unlikely that you will encounter this on a network nowadays. In a lab, it’s simple…we just add the broadcast parameter and our router will forward traffic on all PVCs (emulating broadcast). It might be possible that on a real network, the frame-relay provider would not permit this. That’s why you might have wanted to use the OSPF non-broadcast network type.
Rene
Hi Rene,
So even with not existing connectivity between the SPOKES, is this considered at working scenario ? do we need neighborhood between the SPOKEs to get them exchange their routes ? so how does it work if there is no connectivity ?
thx
Ahmad
Hi Ahmad,
This is a working scenario yes. The two spoke routers establish a full neighbor adjacency with the hub (DR) so they are able to learn all routes. This includes routes that are advertised from spoke routers to the hub.
Rene
Hi Rene
I have my network configured just as you, but I cannot ping Spoke 3 from Spoke 2 , why is that ?
Hi Houari,
There’s a couple of things to check. First of all, do you have an entry in your routing tables of spoke 2+3 so they know how to reach each other?
If so, are they able to reach those next hop IP addresses? If not, you might have to add another frame-relay map on spoke 2 and 3.
Rene
Hi Rene
I don’t see DLCI configured on Hub and spoke routers,
the illustration show DLCI numbers.
please explain
Thank you
Hi,
Is it possible to simulate this lab in gns3 ?
If R1-----R2-------R3 connected like this , Is it a non broadcast type
Thanks
Hello Sims!
The network topology that you are describing would not COMPLETELY simulate the frame relay situation in this lesson. What you are describing would require all communication between R1 and R3 to go through R2 just like in the lesson, however, this also means that the two interfaces on R2 that connect R1 and R3 would have to be on the same subnet, something that is not possible. So you would just be creating two point to point links on different subnets, however, in the frame relay example in the lesson, all three routers are in the same subnet.
What you require is to create a frame relay switch on GNS3. You can find some frame relay examples in GNS3 from Rene’s GNS3 Vault here: http://gns3vault.com/labs/frame-relay/
I hope this has been helpful!
Laz
Hi,
“If you select the non-broadcast network type then OSPF will assume you are running a multi-access network. Couple of key things to remember here:”
can you provide a non multi-access ospf network ? And why it is called non multi access ?
The broadcast and non-broadcast network types require a DR/BDR
So which type does not require a DR/BDR ?
Thanks
Actually who cares about frame-relay. Am I not mistaken this is a dead subject.
These lessons should be tailored for DMVPN now. And yes I realize you can probably relate the two technologies but I have enough in my head to learn and I’m sure DMVPN does not behave exactly like frame-relay all the time. Don’t need frame-relay examples ENOUGH UPDATE THIS STUFF. PLEASE 
Hi,
The question was really not about frame-relay .
It was about multi access ,non broadcast and broadcast ospf .
Thanks
Hello Sims!
Although I’m sure you know most of this, let me start by stating the network types for OSPF and stating an example for each:
Point to Point - a serial connection between two routers
Broadcast - two or more routers on the same subnet connected to an Ethernet switch
The next two types of network are used to accommodate multi access technologies that don’t support broadcast. These technologies include Frame Relay and ATM.
Non Broadcast Multi Access - This is used when you have multiple routers connected on the same subnet over a non-broadcast technology such as frame relay. OSPF must emulate a broadcast by sending multiple unicast packets to all destinations. In this case, a DR must be elected and it must be the HUB router that is the DR. Otherwise OSPF will not function correctly.
Point to Multipoint - This is a network where the topology is the same as in the Non Broadcast Multi Access, however, the HUB router creates multiple point to point connections to each remote router, thus emulating multiple point to point connections. Each point to point connection has its own subnet. A DR/BDR is not necessary in this topology.
So, to specifically answer your questions:
"can you provide a non multi-access ospf network ?"
A non-multi access network is just a point to point network. Multi access just means that there are more than two routers involved in a specific topology.
" And why it is called non multi access ?"
There is no specific term “non multi access” used for OSPF network topologies. Non multi access just means point to point. If you are referring to Non Broadcast, then that just means that you are using a technology that doesn’t allow broadcasts to occur such as Frame Relay.
"The broadcast and non-broadcast network types require a DR/BDR
So which type does not require a DR/BDR ?"
Specifically, out of the four types of networks, the Broadcast network and the Non Broadcast Multi Access require DR/BDRs. The Point to Point and Point to Multipoint do not since the basis of their topologies are point to point connections. Note that in a Broadcast network the DR can be any router based on the DR elections, while in the Non Broadcast Multi Access, the DR must be the HUB router for OSPF to function.
I hope this has been helpful!
Laz
Hi
Why did you choose frame-relay for this example . If we use ip ospf network non-broadcast command in an ethernet broadcast network and get the same reult .
Thanks
Hello Sims
Yes, that may be the case, however, the very purpose of the different types of networks that you can configure for OSPF is to accomodate technologies that function differently such as Frame Relay. So the point is to show how these commands accomodate real technologies rather than just showing the fact that you can make OSPF function differently.
Also, we are trying to approximate real-world applications, and you would never EVER implement non-broadcast OSPF on on an Ethernet network.
I hope this has been helpful!
Laz
Hi Rene,
why there is no concept of DR/BDR in EIGRP, as it is in OSPF ? So in case of shared segments,there wont be any triggered update storm in EIGRP as in OSPF ?
Regards,
Nanu
Hello Nanu
It all has to do with the way OSPF as a link-state protocol functions as compared with EIGRP which is a distance vector protocol.
OSPF advertises a large amount of topological information about the network. Each router must learn what every metric is for every link in the OSPF area. When LSAs are exchanged, each router shares with other routers their perspective of the WHOLE network. EIGRP routers on the other hand share information about a portion of the network. So, first off, LSAs are much larger than the data exchanged by EIGRP, so a traffic regulator like the DR/BDR is necessary when there are multiple OSPF routers in the same subnet.
Secondly, and probably most importantly, OSPF creates neighbor relationships with ALL OSPF routers in its subnet whereas EIGRP creates neighbor relationships only with DIRECTLY CONNECTED EIGRP routers. So information exchange between EIGRP routers is drastically reduced as compared to OSPF.
I hope this has been helpful!
Laz
Hello Laz,
Thank you so much for the explanation, but I still have following doubts:
Regards,
Nanu
Hello Nanu
Yes, I understand your concern. Keep in mind that EIRGP also has a split horizon rule, which states that any information that is learned via an interface is never advertised via that interface. This rule will prevent any “flooding” of EIGRP packets within a segment.
I hope this has been helpful!
Laz
Hello Laz,
Thank you so much for your guidance, this absolutely cleared my doubt.
Regards,
Nanu