Introduction to OSPF

(Shantel - split this topic #39

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


(Rene Molenaar) #40

Hi Anand,

One of the major differences between BGP and any of the IGPs (RIP, OSPF, EIGRP, IS-IS) is that BGP always uses unicast. OSPF for example uses multicast on multi-access networks and has autodiscovery of neighbors.



(Rosa S) #41

The way that you teach is amazing.




(Roshan K) #42

Hi Rene,

After Cisco IOS Ver 15.0 maximum paths OSPF/EIGRP can support upto has been increased to 32 paths I think. Please correct me if I am wrong, I checked this out on GNS3 7200 series router.


(Rene Molenaar) #43

Hi Roshan,

I didn’t know that but that sounds right:

Router(config)#router eigrp 1
Router(config-router)#maximum-paths ?
<1-32> Number of paths

Router(config)#router ospf 1
Router(config-router)#maximum-paths ?
<1-32> Number of paths

Router(config)#router rip
Router(config-router)#maximum-paths ?
<1-32> Number of paths

Router#show version
Cisco IOS Software, 2800 Software (C2800NM-ADVENTERPRISEK9-M), Version 15.1(4)M10, RELEASE SOFTWARE (fc2)

Thanks for sharing!



(Hussein Samir) #44

Hi Rene,

I have several questions :-

1 - Can we have ABR between sub areas or the ABR must be only between backbone area and sub areas ?
2 - Can we have the ASBR in sub areas or the ASBR must be only in the backbone area ?
3 - What is the different between saves address space and used address space that you mentioned in the advertisement of loopback interface addresses ? I did not understand that and I have some ambiguity at this point ? can you explain it to me in detail please ?
4 - what is the benefit of sharing information with OSPF hello packets like Router ID, neighbors, DR and BDR IP address, router priority ?



Hussein Smaeer


(Rene Molenaar) #45

Hi Hussein,

  1. You can't have an ABR between two areas since all areas have to be directly connected to the backbone. It is possible to have an ABR that has an interface in the backbone area and two different areas but it won't be an ABR between area 1 and 2.
  2. You can have an ASBR in any of the areas except for stub and totally stub areas.
  3. OSPF prefers to use the IP address on a loopback interface as its router ID. We can choose if we advertise the loopback interface in OSPF or not, that's it.
  4. Some of the items in the hello packet have to match before we become neighbors, like the area number and area type (stub, totally stub, nssa, totally nssa). Other things are used during the establishment of the neighbor adjacency. We need to figure out which router will become DR/BDR, who starts the exchange of LSAs first, etc.

(Hussein Samir) #46

Thanks Rene,

Can we change the backbone area or the backbone area must be always area 0 ?


(Rene Molenaar) #47

Hi Hussein,

Area 0 is the backbone area by design, you can’t change it.

If you run single area OSPF then it’s no problem to use another area number though.



(Praveen Kumar K) #48

Hi Rene,

Very nice intro to OSPF. I have one question:

we know that the ABR floods a type 3 LSA for every router LSA or network LSA from one area to the other.

type 3 LSA just has the network address info. how do other details like different interfaces connected to the router be advertised accross to other areas.

I get this doubt since each router LSA can have multiple links and how all these are advertised by the ABR

not sure if I have made some sense in that question :slight_smile:


(Rene Molenaar) #49

Hi Praveen,

You won’t find this in LSA type 3. The only thing it carries is the network and the advertising router, that’s it. Keep in mind that OSPF behaves like a distance vector protocol between areas. We only know the topology of our own area, not from other areas.



(Davis W) #50

Happy new year Rene,

Very good explanation!





Hi Rene,

I have completely read your OSPF sessions and its so easy to learn and made me interest to learn all others topics as well. Thanks a lot for your lessons.

I have some queries. Please clear me.

OSPF normally use area 0 as backbone. Please help me in knowing the reason behind this design.

I have read in some book that, one of the reason for designing area0 as backbone in ospf is for loop prevention.

Can you please explain 1) How area 0 concept is helpful to avoid loops inside ospf network.

  1. What all other benefits of area0 methodology.

(Rene Molenaar) #52

Hi Udayakumar,

OSPF is a link-state routing protocol which means that we build a topology database. Each router will know what the topology looks like.

However, this topology database is built PER area. A router in area 0 will know what the topology looks like for area 0 but doesn’t know what other areas look like.

Between areas, OSPF works similar like a distance vector routing protocol (RIP or EIGRP). To ensure that there is a loop-free topology, we have a hierarchical design where all areas have to be connected to a single area (the backbone area). When you go want to go from area 1 to area 2, you can’t go directly from one area to another…it always has to go through the backbone area.



(Shivam K) #53

Hello Rene,

If a router running OSPF is learning two different cost routes for a destination and obviously only one will be installed in Routing table.

What command can display both all the routes for a destination learnt by OSPF with all the details i.e Cost, Metric etc.

Please comment.



(Andrew P) #54

OSPF is different than distance vector routing protocols. The job of OSPF is not to advertise routes. Instead, its job is to solve the shortest path tree to all nodes in its area (OSPF becomes more distance vector-like with multiple areas), but I am referring to just a single area for now. The networks attached to OSPF nodes are learned as attributes of the nodes–they are not learned or shared directly.

What all this means is that there really isn’t a clear command to show all routes learned by OSPF (at least nothing like show ip eigrp topology all-links). Instead, you can ask to see about all nodes and what their associated network attributes are. This is done via
#show ip ospf database

You can also request to view more detailed information on a specific part of the OSPF database depending on what LSA Type you are interested in. This is where it is very important to understand the difference between LSA Types 1, 2, 3, 4, 5, and 7. If you do a
#show ip ospf database ?
You will see all of your options for the portions of the database for more information. The most important ones are:
Router (LSA Type 1)
Network (LSA Type 2)
Summary (LSA Type 3)
ASBR-Summary (LSA Type 4)
External (LSA Type 5)

So, for example, if you wanted to see all the OSPF database information about all destinations within your same OSPF area, you would issue:
#show ip ospf database router

Likewise, to see all information about outside networks being advertised into OSPF, issue:
#show ip ospf database external



(Rahul S) #55

Hello Rene,

I have a question regarding cost .

u said that cost is referece BW/ actual if we have RF BW as 100 mbps and have a FA ethernet which has a maximum capacity of 100 mbps ,but we are configuring it with 80 mbps BW ,that how shd we calculate the cost .

A.100/100 equal cost 1

B.100/80 equal cost 1.25 which is believe is not the correct cost .

so my doubt is that the divided BW would be the maximum bandwidth capability of the interface such as ethernet -10 mbps , FA -100 mbps, Gi -1000mbps ,10gi-10000 mbps.


(Rene Molenaar) #56

Hi Rahul,

The reference bandwidth is a global value and by default, it’s 100Mbit:

R1#show ip ospf | include Ref
 Reference bandwidth unit is 100 mbps

So by default, a 100Mbit interface will have a cost of 1 since 100/100 = 1. When you set the interface to 80 Mbps then the cost should be 100 / 80 = 1.25

The problem, however, is that OSPF only uses rounded values so in this case, 1.25 becomes a cost of 1:

R1(config)#interface GigabitEthernet 0/1
R1(config-if)#bandwidth 80000 
R1#show ip ospf interface GigabitEthernet 0/1 | include Cost
  Process ID 1, Router ID, Network Type BROADCAST, Cost: 1

In other words, OSPF will assign the same cost to an interface with 100 Mbps or 80 Mbps bandwidth. To prevent this from happening, it’s better to set the reference bandwidth to Gigabit (or higher):

R1(config)#router ospf 1                     
R1(config-router)#auto-cost reference-bandwidth 1000
R1#show ip ospf interface GigabitEthernet 0/1 | include Cost
  Process ID 1, Router ID, Network Type BROADCAST, Cost: 12

Hope this helps.



(William V) #57


Hope you are well … can a single router which is an ABR have connections as well as being a member to MORE THANK 1 area including area 0 as one of the member areas ? I am thinking not …

Many Thanks Will

Reason I ask is that I see examples using virtual links where the ABR or area say cannot reach area 0 or get a connection to area 0 , but I cannot find an example in terms of a network diagram where an ABR is connected to more than one area including area 0

Many thanks as always champion


(Shantel - split this topic #58

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