OSPF (1)
Last updated
Last updated
OSPF is a Link State dynamic routing protocol.
[[Dynamic routing#Types#Distance vector]]
Use ‘routing by rumor’ in which each router shares information about the routes they know and their metric cost to reach each destination.
Routers don’t have a complete map of the network, they just use the information their neighboring routers tell them to determine the best route to each destination.
[[Dynamic routing#Types#Link state]]
Every router creates a ‘connectivity map’ of the network.
To allow this, each router advertises information about its interfaces (its connected networks) to its neighbors.
These advertisements are passed along to other routers, until all routers in the network develop the same map of the network.
Then, each router independently uses this map to calculate the best routes to each destination.
Link state protocols use more resources on the router, because more information is shared.
Link state protocols tend to be faster in reacting to changes in the network.
Open Shortest Path First.
The OSPF protocol uses the ‘shortest path first’ algorithm, created by Dutch computer scientist Edsger Dijkstra.
Another name for the algorithm is ‘Dijsktra’s algorithm’
3 versions :
OSPF v1 (1989) : old & not in use
OSPF v2 (1998) : used for IPv4 (in CCNA)
OSPF v3 (2008) : used for IPv6
Routers store information about the network in LSA (Link State Advertisements), which are organized in a structure called the LSDB (Link State Database).
Routers will flood LSAs until all routers in the OSPF area develop the same map of the network, meaning the same LSDB.
3 main steps:
Become neighbors with other routers connected to the same segment.
Exchange LSAs with neighbor routers.
Calculate the best routes to each destination and insert them into the routing table.
![[lsa flooding.png]]
Say this network of four routers is running OSPF. All of these routers are OSPF neighbors, they have the same link state database and the network is stable.
Then, OSPF is enabled on R4’s G1/0 interface.
So, R4 needs to tell the other routers about this new network segment.
So, R4 creates an LSA to tell its neighbors about the network on g1/0. Some basic info in LSA :
RID (router ID). R4’s router ID is 4.4.4.4, either manually configured or one of its loopback interfaces, since there is no physical interface with that IP.
IP : 198.168.4.0/24, network on the g1/0 interface.
Cost : 1 (OSPF's metric).
Each individual LSA has an aging timer, which is 30 minutes by default.
The LSA is then flooded throughout the network (blue arrows) until all routers receive a copy.
The LSA will be flooded again after the timer expires (once every 30 minutes by default).
This results in all routers in the OSPF area having the same LSDB.
The LSDB contains LSAs for all of the different links in the network.
Now that OSPF has been activated on R4’s G1/0 interface, that new LSA is added to the LSDB.
LSDB is identical for all routers in the OSPF area.
Each router then uses the SPF algorithm, Dijkstra’s algorithm, to calculate its best route to 192.168.4.0/24.
Each of these routers has a complete map of the network. For example, R2 is basically looking at the same diagram. so it can calculate that sending traffic via g1/0 is the best route (red arrow).
OSPF uses areas to divide up the network.
Small networks can be single-area without any negative effects on network performance.
In larger networks (network with 500 routers with over 1000 subnets), a single-area design can have some negative effects.
SPF algorithm takes more time to calculate routes in a large network.
Requires exponentially more processing power on each router to make calculations.
Each router sharing a single, huge link state database also takes up more memory on the routers.
Every small change on the network, for example a new interface being activated, would caused LSAs to be flooded to all 500 routers, and all of those routers would have to do the SPF calculation again.
By dividing a large OSPF network into several smaller areas, you can avoid the above negative effects.
![[ospf areas.png]]
An area is a set of routers and links that share the same LSDB.
There are four areas in the above diagram.
Each of these areas maintains a unique LSDB.
The backbone area (area 0) is a special area, that all other areas must connect to.
Notice that area 1, area 2, and area 3 all connect to area 0, the backbone area.
Say, area 1 is not connected to area 0 (backbone area) and only connected to area 2. This network design will not be allowed in OSPF.
![[ospf internal routes.png]]
Routers with all interfaces in the same area are called internal routers.
![[ospf abr.png]]
Routers with interfaces in multiple areas are called area border routers (ABR).
ABRs maintain a separate LSDB for each area they are connected to.
It is recommend that you connect an ABR to a maximum of 2 areas.
Connecting an ABR to 3+ areas can overburden the router.
Shown above is good OSPF network design, with each ABR only connected to 2 areas.
![[ospf backbone routers.png]]
Routers with atleast one interface connected to the backbone area (area 0) are called backbone routers (also includes area border routers).
Intra-area route is a route to a destination inside the same OSPF area.
Destination is in the same area as the router.
Router in area 1 to a destination that is also in area 1.
Interarea route is a route to a destination in a different OSPF area.
Router in area 1 learns a route to a destination in area 2.
Router and the destination are in two different OSPF areas.
![[ospf non contiguous.png]]
OSPF areas should be contiguous.
It means that each individual area should be connected, not divided up.
The network design above is not allowed in OSPF and will cause problems.
So, instead of having area 1 split up and non-contiguous, you should make the section on the right a separate area, area 3.
![[ospf abr to backbone.png]]
All OSPF areas must have at least one ABR connected to the backbone area.
Area 1 does not have an ABR connected to the backbone area, area 0 -> not allowed and will cause problems.
Area 2 has an ABR connected to area 2 and 0, and Area 3 also has an ABR connected to area 3 and 0 -> good network design.
![[ospf subnet.png]]
OSPF interfaces in the same subnet must be in the same area.
If they’re not in the same area, they won’t become OSPF neighbors and won’t exchange information about the networks.
Even though all 4 interfaces are in the same subnet and OSPF is enabled on them, area 1 router will not become OSPF neighbors with the others.
Area 1 router also has an interface in the 192.168.1.0/29 subnet, but the interface is in area 1, not area 0.
To make it an OSPF neighbor, configure ABR’s interface in the 192.168.1.0/29 subnet in area 0.
![[ospf config.png]]
All of these router interfaces are in OSPF area 0.
R1(config)# router ospf process id R1(config)# router ospf 1
Enter OSPF configuration mode you use the command ROUTER OSPF, followed by a process ID.
A router can run multiple OSPF processes at once, and this ID is used in the router to identify each of them.
Typically you’ll just use a single OSPF process.
For EIGRP routers to become neighbors, their AS number has to match.
The OSPF process ID is locally significant.
Routers with different process IDs can become OSPF neighbors. You could use process ID 1 on R1 and process ID 2 on R2, and they would still become OSPF neighbors and exchange LSAs.
R1(config-router)# network 10.0.12.0 0.0.0.3 area 0
The network commands tells OSPF to look for any interfaces with an IP address in the range specified, and then activate OSPF on the interface in the specified area.
OSPF also uses wildcard masks like in EIGRP.
For single area OSPF it’s actually possible to use any area number, but it’s considered best practice to use area 0.
When OSPF is activated on the interface, the router will try to become OSPF neighbors with other OSPF-activated neighbor routers (R1 will become OSPF neighbors with R2 and R3).
R1(config-router)# passive-interface g2/0
The passive-interface command tells the router to stop sending OSPF ‘hello’ messages out of the interface.
However, the router will continue to send LSAs informing its neighbors about the subnet configured on the interface.
So, although R1 won’t send hellos out of g2/0 and try to find OSPF neighbors, it will still tell its other neighbors about the 172.16.1.0/28 network.
Routers will send OSPF hello messages on loopback interfaces, even though they are connected to nothing. So make the loopback interface a passive interface too.
Always use this command on interfaces which don’t have any OSPF neighbors.
R1(config)# ip route 0.0.0.0 0.0.0.0 203.0.113.2
Configured a default route on R1, and the next hop is the ISP’s IP address.
R1(config-router)# default-information originate
This will cause the router to create a new LSA and flood it.
You can see R2, R3, R4 added the default route via R1 to its route table.
Configuring a default route and then advertising it into OSPF using this command will make R1 an OSPF ASBR (Autonomous System Boundary Router).
R1(config-router)# router-id ID in IP format R1(config-router)# router-id 1.1.1.1
In EIGRP the command is 'eigrp router-id', in OSPF it's just 'routere-id'.
After you enter this, the router says to reload or use ‘clear ip ospf process’ command, for this to take effect.
R1# clear ip ospf process Reset ALL OSPF processes? {no} : yes
This basically resets OSPF on the router.
Bad idea in a real network, and the router will lose all of its OSPF routes for a short time and won’t be able to forward traffic.
'no' in square brackets indicates the default choice. If you press 'enter' the router will assume 'no'.
R1(config-router)# maximum-paths <1-32>
Default value is 4.
R1(config-router)# distance <1-225>
If you change it to 85, OSPF routes are preferred over EIGRP routes on this router.
R1# show ip protocols
At the top it says ‘routing protocol is ospf 1’. 1 is the process ID configured earlier.
Then the router ID. Same as EIGRP. Order of priority :
Manual configuration
Highest IP address on a loopback interface
Highest IP address on a physical interface
Then it shows ‘It is an autonomous system boundary router’.
ASBR is an OSPF router that connects the OSPF network to an external network.
R1 is connected to the Internet. By using the 'default-information originate' command, R1 becomes an ASBR, it connects the OSPF network to the Internet.
You won't see this in other routers as they are not connected to external networks.
Next, 'number of areas in this router is 1. 1 normal 0 stub 0 nssa'.
These are three different types of OSPF areas.
Number of areas this router is in, is one, because this is single-area OSPF.
Next, maximum paths is 4.
Unlike EIGRP, OSPF doesn’t support unequal-cost load-balancing.
It does support ECMP load-balancing over a maximum of 4 paths by default and can be changed.
The ‘routing for networks’ section shows the network commands entered.
This only determines which interfaces OSPF will be activated on, it doesn’t tell OSPF to flood LSAs for these specific networks.
Passive interfaces.
R1's neighbors.
OSPF’s AD (default 110).
![[ospf 1 quiz.png]]
OSPF will advertise RouterA's gateway of last resort, that is the default route.
An ASBR is an OSPF router that is connected to an external network, outside of the OSPF domain and it shares information into the OSPF domain about those external networks.