The process of breaking up a single area into multiple smaller areas is also known as hierarchical design. OSPF can operate well in larger environments with a hierarchical routing design in which multiple areas span or link to the core backbone Area 0. In addition to the benefits mentioned in the introduction, network problems such as flapping can be isolated to smaller areas without an adverse effect on the backbone area due to excessive SPF recalculations. When designing an OSPF network, you need to establish which routers and links are to be included in the backbone and which ones are to be placed in each area. As you know, an area is a logical grouping of network segments that consists of a single autonomous system (AS). An area can also be viewed as a subset of an AS that is defined to reduce the number of routing updates and SPF calculations. In addition, confining link-state updates to defined areas of the internetwork can reduce convergence time. There are several different types of areas that are used to manage link-state information propagation. We will explore these in greater detail after you have learned about the varying network and link-state data types. OSPF Router TypesThe area ID 0 must be used to represent the backbone area. A backbone router (BBR) is characterized as a router that has at least one interface connected to Area 0 (or OSPF transit area). Therefore, Area 0 is commonly referred to as the backbone area . A backbone is simply the portion of a network that functions as the primary path for source traffic to flow to other networks. The backbone can also be a collection of OSPF routers acting as a conduit for information moving between different OSPF autonomous systems. Internal routers (IRs) have all their interfaces connected to a single OSPF area. Therefore, internal routers in a single area share the same link-state databases. Internal routers can exist in a backbone, a standard area, or a stub area. Standard and stub areas are defined later in this chapter. By definition, routers that have interfaces linked to multiple areas and that contain separate link-state databases for each area are called area border routers (ABRs). Practically speaking, ABRs are the routers that connect an area to the backbone. An area can have one or more ABRs. Area border routers serve as the gateways of last resort for internal routers that are sending information to external routes. They are the principal exit points for the area, so routing data meant for another area must travel through the local ABR to get there. Areas should have a contiguous set of network (and/or subnet) addresses so that they can deploy route summarization techniques for smaller routing tables. It is preferable to have more than one ABR per area for fault tolerance and performance reasons. Internal routers have identical link-state databases, whereas ABRs maintain a separate link-state database for each area in the router's autonomous system.
An autonomous system boundary router (ASBR) is an OSPF router with at least one interface connected to an external non-OSPF EGP network, another IGP, or between different OSPF autonomous systems. These routers facilitate the transport of information between the IGP (OSPF is the IGP used in this scenario) and the border gateway protocols (BGPs). The ASBR bridges the gap between the OSPF interface and the BGP-bound link by exchanging routing table data with the BGP.
Table 5.1 displays the four main types of OSPF routers that can participate in an autonomous system, and Figure 5.1 illustrates all the router types. Figure 5.1. Example of multiple OSPF router types.
Table 5.1. Types of Routers
In Figure 5.1, the ABRs connecting Area 0 to Areas 1 and 2 can both be categorized as backbone routers as well as ABRs. If the backbone router inside Area 0 were connected to another external AS, it would be defined as an ASBR and backbone router. In a multi-area OSPF network, Area 0 (the backbone) collects link-state advertisements from all areas of the internetwork. Every area must either be physically or logically connected to Area 0. For areas that are not physically connected to the backbone area, OSPF includes the concept of virtual (logical) links. A virtual link generates a logical path that transits between two ABRs, a backbone ABR, and another area that has no physical link to Area 0. Although this mechanism is provided, it is not recommended that you design an OSPF network to require virtual links. The use of virtual links is a stopgap measure for the short term until a better solution is implemented. Because the stability of a virtual link is dependent upon the stability of the underlying area, this can make troubleshooting more difficult. In addition, virtual links cannot run across stub areas. You can use the area virtual-link router configuration command on the ABRs that reside on either side of the area that is being traversed (this can also be referred to as a transit area). Here is the official syntax for the command: RouterA(config-router)# area area-id virtual-link router-id The area-id parameter is the area identifier for the transit area. The router-id parameter is the router for the virtual link neighbor. To remove a virtual link, use the following router configuration command: RouterA(config-router)# no area area-id virtual-link In Figure 5.2, the area virtual-link router configuration command is used to create a logical connection between Area 2 and the backbone Area 0. Although the backbone area is usually configured as the transit area, Area 1 would be considered the transit in this example. Figure 5.2. Creating a virtual link.
Here is the correct configuration for RouterA and RouterB to establish a virtual link between Area 0 and Area 2: RouterA(config)# router ospf 50 RouterA(config-router)# network 172.16.0.0 0.0.255.255 area 0 RouterA(config-router)# network 172.17.0.0 0.0.255.255 area 1 RouterA(config-router)# area 1 virtual-link 172.18.1.2 RouterA(config)# router ospf 70 RouterA(config-router)# network 172.17.0.0 0.0.255.255 area 1 RouterA(config-router)# network 172.18.0.0 0.0.255.255 area 2 RouterA(config-router)# area 1 virtual-link 172.17.1.1 The router ID value is derived from the highest IP address value that is configured for the router's active interfaces. Administrators often configure an IP address as a loopback address, so that the higher loopback address becomes the router ID. A loopback address is a virtual interface that always remains active and will not fail as a physical interface is known to do. To configure a loopback address on a router, use the following syntax: RouterA(config)# interface loopback 1 RouterA(config-if) ip address 192.168.200.254 255.255.255.0 You can then publish a loopback address as you would advertise any network with the network area command, as follows : RouterA(config)# router ospf 10 RouterA(config-router)# network 192.168.200.254 0.0.0.0 area 0 Remember, an unpublished loopback address cannot be pinged. You have now established a static router ID that is not dependent on the status of a particular active interface using a loopback address. Layer 3 addressing remains active for OSPF router operations even if an individual physical interface goes down. OSPF Link-State Area TypesWe have determined that an area is a subset of an autonomous system. However, there are various types of routing updates that can be propagated depending on the environment. To control the number of SPF activities, diminish the size of tables, and reduce LSU overhead, areas can be defined according to the following descriptions:
OSPF Link-State Advertisement TypesBy default, OSPF floods link-state updates (LSUs) to every adjacent neighbor every 30 minutes to make certain that each router in the OSPF routing domain has a current link-state database. These neighbors then update their own link-state databases and continue flooding the link-state change to their adjacent neighbors.
When an OSPF router senses a change in the topology, it sends the change information via an LSA combined with an LSU packet to the multicast address 224.0.0.6. OSPF uses only multicast ”not broadcast ”addresses when flooding link-state modifications throughout a network. This action tells the DR and BDR that a topology change has occurred. After acknowledging receipt of the packet to the sourcing router, the DR uses the multicast address 224.0.0.5 to flood the LSU through all its active interfaces. Any OSPF router listening to this address will be notified of the change, acknowledge receiving the LSU, and continue flooding the LSU until convergence occurs. OSPF routers send a complete copy of their topology database only when initially establishing adjacency with a neighbor router. Table 5.2 explains the six different types of link-state advertisements supported by Cisco routers and a detailed description of how they are generated, as well as which area types they are associated with. This is an important table to memorize for a testing scenario. Compare the following table with Figure 5.3 to visualize how the different LSA types are implemented. Figure 5.3. Different OSPF LSA types are used intra-area and inter-area.
Table 5.2. Types of Link-State Advertisements (LSAs) Supported by Cisco Routers
|