Open Shortest Path First (OSPF)

Open Shortest Path First (OSPF) differs from IGRP and Enhanced IGRP (EIGRP) in that it is a pure link-state routing technology. Link-state routing protocols only send updates when the topology of the network changes. Also, OSPF is an open standard routing protocol, which means that it was not developed solely by Cisco. OSPF was designed and developed by the IETF to provide a scalable, quickly converging, and efficient routing protocol that could be used by all routing equipment. Complete details for OSPF are found in RFC 1247. You can implement several types of OSPF areas—each of these types will be discussed in detail. Configuration examples and information specific to OSPF such as metrics, redistribution, and filtering will be explained.

OSPF supports several features that RIP version 1 does not; for example, VLSM and route summarization are supported by OSPF but not by RIP.

The hop count was eliminated with OSPF, thus giving it limitless reachability. RIP was limited to 16 hops. Due to the algorithm used to calculate and advertise routes, network convergence, or the time it takes for the network routers to update their route tables, is fast with OSPF. OSPF is like EIGRP in that it sends route updates only when changes occur in the network. A formal neighbor relationship is established with all adjacent OSPF routers.

Areas are used within OSPF to define a group of routers and networks belonging to the same OSPF session. Links connect routers, and the information about each link is defined by its link state. On each broadcast or multi-access network segment, two routers must be assigned the responsibilities of designated router (DR) and backup designated router (BDR). Designated routers are OSPF routers that generate link-state advertisements (LSAs) for a multi-access network and have other special responsibilities in running OSPF. Each multi-access OSPF network that has at least two attached routers has a designated router that is elected by the OSPF Hello protocol. The designated router enables a reduction in the number of adjacencies required on a multi-access network, which in turn reduces the amount of routing protocol traffic and the size of the topology database.

Like EIGRP, OSPF maintains three databases: adjacency, topology, and route. The adjacency database is similar to the neighbor database used by EIGRP. It contains all information about OSPF neighbors and the links connecting them. The topology database maintains all route information. The best routes from the topology database are placed in the route database, also known as the route table.

That was a billboard summary of OSPF. Now that you have a general understanding of how OSPF is set up, let’s move on and discuss its features and operation in detail.

OSPF Features and Operation

OSPF operation can be divided into several categories. Let’s start with how OSPF initializes and creates a peering relationship with adjacent routers. After an OSPF area is defined, we will discuss how the DR and BDR are chosen and what their responsibilities are. Once the area is established and functioning, routing information must be learned and shared.

Initializing OSPF

The Hello protocol is used to establish peering sessions between routers. Hello packets are multicast out every interface. The information that is multicast includes

• The router ID

• Timing intervals

• Existing neighbors

• Area identification

• Router priority

• Designated and backup router information

• Authentication password

• Stub area information

All this information is used when establishing new peers. Table 9.1 contains descriptions of each of these elements.

Table 9.1: OSPF Multicast Information (continued)

Information	Description
Router ID	The highest active IP address on the router or the loopback interface IP address (even ifitis lower).
Time intervals	Contains intervals between Hello packets and the dead time interval.
Existing neighbors	Addresses for any existing OSPF neighbors.
Area identification	OSPF area number and link information must be the same for a peering session to be established.
Router priority	This value is used when choosing the DR and BDR.
DR and BDR	If they have already been chosen, their information is contained in the Hello packet.
Authentication password	All peers must have the same authentication password if authentication is enabled.
Stub area flag	This is a special areatwo routers must share the same stub information. This is notnecessary to initiate a regular peering session with another OSPF router.

Figure 9.1 displays a flowchart that depicts each step of the initialization process. The process starts by sending out Hello packets. Every listening router will then add the originating router to the adjacency database. The responding routers will reply with all of their Hello information so that the originating router can add them to its adjacency table.

Figure 9.1: OSPF peer initialization

Once adjacencies have been established, the DR and BDR need to be chosen before route information and link-state information can be exchanged. After the DR and BDR are chosen, route information is exchanged, and the OSPF peers continue to multicast Hello packets every 10 seconds to determine if neighbors are still reachable.

Before proceeding further with peer initialization, let’s discuss several terms specific to OSPF. These terms are important for understanding OSPF and how it functions.

The easiest way to understand OSPF is to build from what you already know about EIGRP. In Chapter 8 you learned that EIGRP uses autonomous system (AS) numbers to specify routing processes and the routing process to which individual routers belong. OSPF uses areas in addition to autonomous systems. OSPF areas participate within a single AS and consist of a group of routers or interfaces on a router that are assigned to a common area.

OSPF also allows and uses different area types. When deploying OSPF, there must be a backbone area. The standard and stub areas connect to the backbone area. Here is a list of each router type, followed by a short description of the area type:

Backbone The backbone area accepts all LSAs and is used to connect multiple areas.

Stub The stub area will not accept any external routing update, but it will accept summary LSAs.

Totally stub The totally stub areas are closed off from accepting external or summary advertisements.

Standard The standard area is the normal area that accepts internal and external LSAs and summary information.

From the previous area descriptions, you probably noticed the frequent use of the term LSA. LSAs are the heart of OSPF’s information exchange. LSAs distribute a list of active links to neighbor routers. Different types of LSAs represent different types of route information. Table 9.2 summarizes all of the LSA types.

Table 9.2: OSPF LSA Types

LSA Type	Description
1. Router link entry	This LSA is broadcast only within its defined area. The Type 1 LSA contains all of the default link-state information.
2. Network entry	This LSA is multicast to all area routers by the DR. This update contains network-specific information.
3./4. Summary entry	Type 3 LSAs contain route information for internal networks and are sent to backbone routers. Type 4 LSAs contain information about Autonomous System Boundary Routers (ASBRs). Summary infor mation is multicast by the Area Border Router (ABR), and the information reaches all backbone routers.
5. Autonomous system entry	As the name indicates, these advertisements originate from the ASBR. These packets contain information about external networks.

Different LSA types represent the type of route that is being advertised and assist in restricting the number and types of routes that are accepted by a given area. As shown in Table 9.2, an LSA of Type 5 is sent only by the ASBR.

Multiple router types can exist within an OSPF area. Table 9.3 lists all of the OSPF router types and the role that each plays within the area.

Table 9.3: OSPF Router Types

Router Type	Description
Internal	All interfaces are defined on the same area. All internal routers have an identical link-state database.
Backbone	Backbone routers have at least one interface assigned to area 0.
Area border router (ABR)	Interfaces are connected to multiple OSPF areas. Information specific to each area is storedon ABRs.
Autonomous system boundary router (ASBR)	An ASBR has an interface connected to an external network or to a different AS.

In addition to the listed responsibilities, a router can also be assigned additional responsibilities. These responsibilities are assumed when a router is assigned the role of DR or BDR.

Designated and Backup Designated Routers

The DR and BDR are focal points within each multi-access network segment. These special routers act something like switchboards for information. Routers on any given segment will exchange link-state information with the DR and BDR instead of with each other. After a router has updated the DR and BDR with its information, the DR or BDR multicasts this information to all other routers on this segment. The DR and BDR are also in charge of making sure that all routers have the same version of link-state information. The BDR listens to all updates, but it sends OSPF information on a segment only when the DR fails.

Once a DR and BDR have been established, new routers will establish adjacencies only with the DR and BDR.

Initializing and Maintaining Route Information

Routes are discovered via the Exchange Protocol, as illustrated in Figure 9.1. The Exchange Protocol commences only after the DR and BDR have been chosen. The Exstart state depicted in Figure 9.1 creates peering relationships with the DR, BDR, and each individual router within the area.

The following steps are taken when route information is exchanged (also known as database synchronization):

1. Master-slave relationships are established between routers; the router with the highest router ID is the master. And just like master-slave relationships of the human kind, the slave does most of the work.

2. Master and slaves exchange database description packets (called DBDs or DDPs). These packets contain LSA information, which consists of the IP address of the advertising router, cost, and sequence number or data that indicates how recent the link-state information is.

3. After the slave router receives the DBD or DDP, it sends back an LSAck (link-state acknowledgment) packet. It also compares new information to the existing information. If the information provided by the DBD is newer than the existing information, an LSR (link-state request) is sent back to the master router. The LSR tells the master router to send complete information. The sending of LSRs occurs in the loading state.

4. The master responds with a link-state update (LSU).

5. The slave replies with another LSAck packet.

6. All routers within the area receive the link-state information via the DR.

7. A router will continue to send LSRs until it receives all the information that it needs. Once it has all of the information it needs, it will change to the full state.

8. Once a router is in full state, it can route traffic. At this point in the exchange, all routers should have the same link-state database.

Look at the example in Figure 9.2. This is a protocol trace from a network. Look at the final exchange between the master sending out the LSU and other routers replying with LSAck packets. Note that the destination address of all the LSUs and LSAck packets are multicast addresses.

Figure 9.2: OSPF LSU and LSAck exchange

Now that you have seen the trace, let’s look at the actual packets. First, you’ll see the LSU packet, and then you’ll see the LSAcks after the LSUs have been sent:

IP Header - Internet Protocol Datagram   Version:              4   Header Length:        5   Precedence:           6   Type of Service:      %000   Unused:               %00   Total Length:         76   Identifier:           14145   Fragmentation Flags:  %000   Fragment Offset:      0   Time To Live:         1   IP Type:              0x59  OSPF   Header Checksum:      0x5ba6   Source IP Address:    131.31.194.141   Dest. IP Address:     224.0.0.5   No Internet Datagram Options OSPF - Open Shortest Path First Routing Protocol   Version:                  2   Type:                     4  Link State Update   Packet Length:            56   Router IP Address:        153.53.193.1   Area ID:                  1   Checksum:                 0x6aa5   Authentication Type:      0  No Authentication   Authentication Data:   ........           00 00 00 00 00 00 00 00   # Of Advertisements:      1 Link State Advertisement Header   Age:                      3600  seconds   Options:                  %00100010         No AS External Link State Advertisements   Type:                     3  Summary Link (IP Network)   ID:                       0x90fb6400   Advertising Router:       153.53.193.1   Sequence Number:          2147483708   Checksum:                 0x3946   Link State Length:        28 Summary Links Advertisement (IP Network)   Network Mask:             0xffffff00   Type Of Service:          0   Metric:                   0xffffff Frame Check Sequence:  0x20536f66 

From this packet decoding, you can see that the IP type is OSPF. From there, skip down to the OSPF header information; it indicates that the packet is an LSU. Also in the packet decoding, you can find the IP address of the advertising router and the LSA header. The LSA includes summary information.

Now let’s look at just one of the acknowledgments to the LSU:

IP Header - Internet Protocol Datagram   Version:              4   Header Length:        5   Precedence:           6   Type of Service:      %000   Unused:               %00   Total Length:         84   Identifier:           1285   Fragmentation Flags:  %000   Fragment Offset:      0   Time To Live:         1   IP Type:              0x59  OSPF   Header Checksum:      0x8dda   Source IP Address:    131.31.194.140   Dest. IP Address:     224.0.0.6   No Internet Datagram Options OSPF - Open Shortest Path First Routing Protocol   Version:                  2   Type:                     5  Link State Acknowledgment   Packet Length:            64   Router IP Address:        142.42.193.1   Area ID:                  1   Checksum:                 0x6699   Authentication Type:      0  No Authentication   Authentication Data:   ........           00 00 00 00 00 00 00 00 Link State Advertisement Header   Age:                      3600  seconds   Options:                  %00100010         No AS External Link State Advertisements   Type:                     3  Summary Link (IP Network)   ID:                       0x90fb6400   Advertising Router:       153.53.193.1   Sequence Number:          2147483708   Checksum:                 0x3946   Link State Length:        28 Link State Advertisement Header   Age:                      3600  seconds   Options:                  %00100010         No AS External Link State Advertisements   Type:                     3  Summary Link (IP Network)   ID:                       0x90fb6400   Advertising Router:       131.31.193.1   Sequence Number:          2147483650   Checksum:                 0x25c0   Link State Length:        28 Frame Check Sequence:  0x00000000 

This decoding shows you that the packet is an LSAck. The IP address is the IP address of the responding router. Other LSAs were also picked up on the trace, because each router responds to the LSU.

Link-state protocols choose their routes differently than distance-vector routing algorithms do. Bandwidth is one of the most important metrics for route selection within OSPF. The Dijkstra algorithm (discussed in Chapter 8) is used to choose the lowest cost link for route selection.

The costs of the local router are added to the cost required to reach the destination. The route with the lowest cost is selected as the primary route. Just as with EIGRP, OSPF can hold six equal-cost routes for each destination. Changes in link-state status change the cost of the link.

Every time a link state changes, LSUs are sent to the DR and BDR. The DR relays the information to all other routers within the segment, causing the route table to be recalculated. You have already seen the trace and decoding for this process during the discussion of the route initialization process. If there are many link-state changes within a short period, network convergence may never be reached. OSPF prevents this problem by using hold-down times. The default hold-down time for route calculation is 10 seconds.

Let’s summarize how route information is calculated and shared between other routers:

1. A link-state change occurs.

2. An LSU packet with the new LSA attached is sent to all DRs. The multicast address 224.0.0.6 is used by a router to send to all DRs and BDRs.

3. The DR then notifies all other adjacent routers by multicasting an LSU on the address 224.0.0.5.

4. Each adjacent router responds with an LSAck.

5. Other networks receive updates via LSUs sent by the other network’s DR.

6. The receiving router updates the link-state database.

7. The SPF algorithm is used to calculate the new route table and to control when the router begins to use the new table.

OSPF Metrics

The metrics associated with OSPF are different from those associated with IGRP and EIGRP. OSPF uses bandwidth as the main metric in selecting a route. The cost is calculated using the bandwidth for the link. The equation is 10⁸ divided by the bandwidth. You may change bandwidth on the individual interface.

The cost is manipulated by changing the value to a number within the range of 1 to 65,535. Since the cost is assigned to each link, the value must be changed on each interface. The command to do this is as follows:

ip ospf cost 

Note

Cisco bases link cost on bandwidth. Other vendors may use other metrics to calculate the link’s cost. In other words, you may have to compare apples to oranges or bananas. When connecting links between routers from different vendors, you may have to adjust the cost to match the other router. Both routers must assign the same cost to the link (in apples, oranges, or bananas) for OSPF to work.

You can configure the OSPF distance with the following command:

distance ospf [external | Intra-area | Inter-area] distance 

This command allows the distance metric to be defined for external OSPF and for intra-area and inter-area routes.

Note

Intra-area and inter-area routes are discussed later in this chapter.

Distance values range from 1 to 255, and the lower the distance, the better. Connected interfaces have a distance of 0.

Other values important to OSPF’s operation are not actually metrics, but they can be configured as well. Values such as the router ID and router priority are important in router initialization and DR and BDR selection. You can change these values with some minor configuration changes.

To change the router priority, use the following command on the desired interface:

ip ospf priority number 

The number can range from 0 to 255; the higher value indicates a higher priority when choosing the DR and BDR for the area.

A loopback interface must be added to the router to change the router ID. The IP address of the loopback can be a private address or a fake address. If the IP address is to be announced, a private IP address should be used. To implement a loopback interface on the router, use the following command:

interface loopback number  ip address A.B.C.D 

You can inject the address used on the loopback interface into OSPF by using the network area command.

Configuring OSPF

The initial configuration for OSPF is simple and straightforward, but for connectivity to be established between routers from different vendors, other configuration commands may have to be issued.

Configuration can be broken down into two groups. The first group is for internal routers, or routers within an area; the second group is for routers that connect different areas.

Internal Routers

When you’re initializing OSPF on a router, the session is defined by a process identification. Unlike EIGRP’s AS number, the process identification number does not have to be uniform across all the routers within the OSPF area. Instead of using the process ID to identify the OSPF area, the networks that are added to the session are assigned to an OSPF area. This means that all networks assigned to a given area make up that area. Okay, I admit that is probably obvious!

To initiate OSPF on a router, the first step is to assign the routing protocol with the process ID. After that is done, networks are added and assigned to the desired OSPF area. The commands to do this are as follows:

router ospf process-id network address wildcard-mask area area-id 

The process-id is an integer from 1 to 65,535; the area-id is an integer from 0 to 4,294,967,295.

Let’s look at a configuration example:

Router_A#conf t Enter configuration commands, one per line.  End with    CNTL/Z. Router_A(config)#router ospf 1 Router_A(config-router)#network 172.16.20.0 0.0.0.255    area 20 Router_A(config-router)#network 10.1.2.0 0.0.0.255 area    20 Router_A(config-router)#^Z Router_A# Router_A#show running-config Building configuration... Current configuration: ! version 11.3 no service password-encryption ! hostname Router_A ! enable password aloha ! interface Ethernet0/0  ip address 172.16.10.1 255.255.255.0 ! interface Serial0/0  ip address 172.16.20.5 255.255.255.252  no ip mroute-cache  no fair-queue ! interface Ethernet0/1  ip address 10.1.2.1 255.255.255.0 ! router ospf 1  network 10.1.2.0 0.0.0.255 area 20  network 172.16.20.0 0.0.0.255 area 20 ! ip classless ! line con 0 line aux 0 line vty 0 4  password aloha  login ! end Router_A# 

This example shows that the OSPF process was defined as one. Two networks were added to OSPF area 20. As you can see, the configuration is quite simple. Again, this configuration is the simplest form of OSPF implementation.

Area Configuration and Wildcard Masks

OSPF configurations can get complicated really fast. It is important to understand the networks your router is attached to before attempting configuration. In this section, we will use a sample router configuration and use the network command with a wildcard mask.

The sample router has three interfaces. Interface Ethernet 0 is connected to 172.16.10.0/24 area 0. Serial 0 is connected to 172.16.4.0/22, and Serial 1 is connected to 172.16.8.0/22, both connected to area 1. Notice that we are using a different mask for the WANs than we are for the LANs. This is a benefit of using OSPF; it allows variable-length subnet masks (VLSMs). The router configuration should then look like this:

Router ospf 1 Network 172.16.10.0 0.0.0.255 area 0 Network 172.16.4.0 0.0.3.255 area 1 Network 172.16.8.0 0.0.3.255 area 1

Notice that the backbone connection is defined first. The wildcard mask was simple since a whole subnet was specified. To figure out a wildcard mask, use all 1s minus the network mask. Here is an example:

255.255.255.255 (all 1s) 255.255.255.0   (the mask) 0  .0  .0  .255 (the answer)

For the serial interfaces connected to area 1, remember that the mask was not 24 bits, but 22 bits. The mask would then be as follows:

255.255.255.255 (all 1s) 255.255.252.0   (our mask) 0  .0  .3  .255 (the answer)

OSPF uses areas to group routers. This area, and the topology of the routers, is hidden from the rest of the AS. This allows routing traffic to be reduced, which saves bandwidth.

An OSPF network must have a backbone, or area 0. All other areas must connect to the backbone. Virtual links can be configured for areas that do not, or cannot, connect to the backbone.

To configure area 0, for example, 0 or 0.0.0.0 say the same thing. After the area number, you apply a wildcard mask, which is the same type of wildcard mask used in access lists. By using wildcard masks, OSPF gives you a granular control over the OSPF routing process.

Consider the following example:

Network 172.16.16.0 0.0.7.255 area 0

This command tells the router that only interfaces addressed with subnets 172.16.16.0 will participate in OSPF area 0. Any other interface on the router that is not part of that subnet range will not be part of area 0. Try that with RIP!

To figure out what wildcard mask to use, minus all 1s from the mask. For example, the network 172.16.0.0 in the preceding example is using a /21 network, or 255.255.248.0. The wildcard mask would be configured like so:

255.255.255.255 (all 1s) 255.255.248.0   (the mask)   0.  0.  7.255 (the answer)

Since 256 – 248 equals 8, then you know the subnets are 8, 16, 24, …, 240. If you place a 16 in the third octet, then you are saying you want to use only subnets 16–23.

Wildcard masks can also be used to reference only a certain number of subnets. In this example,

Network 172.16.64.0 0.0.63.255 area 1 Network 172.16.32.0 0.0.31.255 area 2

the first network command tell the OSPF routing process to only let networks defined in the third octet as 64–127 participate in area 1 routing. The second network command tells the OSPF routing process to only let subnets defined in the third octet as 32–63 participate in area 2 routing.

In the first network command example, only two bits are used in the third octet. To figure this out, take eight binary bits and then count up 63. The result is 00111111. This means that the 0s are used to define the networks, and 1s are wild, or can be anything. The zero bits can be all on or all off, hence the addresses of 0 and 192.

In the second network command example, 31 in binary is 00011111. This leaves three bits to define the networks and five bits to be wild. Since the zero bits can be all on or all off, eight subnets can be defined.

Let’s take a look at one more example. This is important to really understand.

Multi-Area Example

This next example has a router with three interfaces:

• Ethernet 0 (E0) is connected to network 172.16.17.0/20.

• Serial 0 (S0) is connected to network 172.16.9.0/21.

• Serial 1 (S1) is connected to network 172.16.7.0/30.

Here’s how the configuration would look:

Router(config)#router ospf 172 Router(config-router)#exit Router(config)#int e0 Router(config-if)#ip address 172.16.17.1 255.255.240.0 Router(config-if)#no shut Router(config-if)#int s1 Router(config-if)#ip address 172.16.9.1 255.255.248.0 Router(config-if)#no shut Router(config-if)#int s0 Router(config-if)#ip address 172.16.7.1 255.255.252.0 Router(config-if)#no shut Router(config-if)#router ospf 172 Router(config-router)#network 172.16.17.0 0.0.15.255     area 0 Router(config-router)#network 172.16.9.0 0.0.7.255 area 1 Router(config-router)#network 172.16.7.0 0.0.3.255 area 2 

Notice that the first command is the router ospf 172 command. If you do not type this command, then you cannot configure VLSM interfaces; you would get an error from the router.

Notice the wildcard addresses. Let’s work through the wildcard mask of each command:

• Network 172.16.17.0 0.0.15.255 area 0:

255.255.255.255 (all 1s)

255.255.240.0 (the subnet mask)

0.0.15.255 (the answer)

• Network 172.16.9.0 0.0.7.255 area 1:

255.255.255.255 (all 1s)

255.255.248.0 (the subnet mask)

0.0.7.255 (the answer)

• Network 172.16.7.0 0.0.3.255 area 2:

255.255.255.255 (all 1s)

255.255.252.0 (the subnet mask)

0.0.3.255 (the answer)

Route Summarization

Route summarization in OSPF is used for the same reasons that other routing protocols use it. When a route table becomes large, it taxes the router in multiple ways: CPU utilization, and bandwidth and memory consumption.

OSPF supports two types of route summarization: inter-area and external. As with many inter-area functions, the area border router (ABR) is in charge of summarizing inter-area routes. External summarization is exactly that— only external routes are summarized by the autonomous system border router (ASBR).

The summarization commands must be issued on the respective routers. Inter-area summarization must be configured on the ABR with the following command within the OSPF routing session:

area area-id range address mask 

The area-id is the OSPF area number. The address and mask define the range of IP addresses that will be summarized for the specified area. To configure external summarization, the following command must be entered within the OSPF routing session on the ASBR:

summary-address address mask 

Notice that the area ID was not used in the second command. This is because only external routes will be summarized. The address and mask define the range of external IP addresses that will be summarized.

Stub Areas

When the route tables and link-state databases grow too large and can’t be remedied with route summarization alone, the area may be configured as a stub area or a totally stub area.

The difference between a stub area and a normal area is that intra-area and inter-area routes are the only routes allowed inside the stub area. To communicate with networks that are not present in the route table, the stub area relies on a default route of 0.0.0.0. This means that when a router does not have a route to the destination address, it will forward the packet to the ABR from which it learned the default network 0.0.0.0. The ABR will then do the route lookup and forward the packet accordingly.

A stub area contains all routes from its area and other areas that are connected via the backbone and ABRs. It does not contain any external routes.

To further shrink a route table for an area, you can make the route table totally stubby. A totally stub area contains an even smaller route table, because the only routes it knows are the routes from within its own area. To contact networks outside the area, the default route 0.0.0.0 is used as well. The ABR is responsible for forwarding packets to the correct destination.

Configuring Stub Areas

To configure a stub area, the following command should be implemented on an internal router that is not an ASBR and not connected to the backbone. It is also important to issue this command on every router within the stub or totally stub area:

area area-id stub

The command is self-explanatory—the area-id simply defines the stub area.

To make an area totally stubby, the same command is used, with one modification. It must be issued on the ABR for the area:

area area-id stub no-summary

Note

You cannot configure routers as stub routers if they belong to the backbone. An area cannot be defined as a stub area if an ASBR is part of the area.

Redistribution

OSPF supports redistribution (the movement of networks from one protocol into another protocol) for many protocols. The following listing directly from a router lists the protocols supported:

Router_A(config-router)#redistribute ?   bgp        Border Gateway Protocol (BGP)   connected  Connected   egp        Exterior Gateway Protocol (EGP)   eigrp      Enhanced Interior Gateway Routing Protocol     (EIGRP)   igrp       Interior Gateway Routing Protocol (IGRP)   isis       ISO IS-IS   iso-igrp   IGRP for OSI networks   mobile     Mobile routes   odr        On Demand stub Routes   ospf       Open Shortest Path First (OSPF)   rip        Routing Information Protocol (RIP)   static     Static routes 

Just as with EIGRP, new metrics must be assigned to route information that is injected into the OSPF session. The command to do this is much simpler than the command used when assigning metrics for EIGRP or IGRP— it is almost the same, but only one metric is assigned. The value of the metric is the cost for the route:

default-metric cost 

OSPF Filtering Considerations

Normal methods of route filtering done by distribution lists are effective when filtering route information. The most effective method of filtering OSPF is to implement the filters on the ASBR as outbound filters. Inbound filters are effective in filtering routes, but since they are inbound filters, LSA packets are still propagated.

Cisco recommends that you filter within other protocols and that you don’t filter OSPF if possible. This is accomplished by implementing outbound filters on other protocols, which keeps unwanted networks from even entering the OSPF area.