Cisco Routing 3 - OSPF - Open Shortest Path First

This entire series on routing was created when I was preparing for the Cisco 642-901 BSCI test, as my notes. I then made a simple edit and published the text. If you think something essential is missing, something is not described correctly or is not entirely clear, I will welcome your information in the comments.

OSPF Properties

• an open standard, probably the most widely used IGP protocol in larger companies
• it is a classless protocol / supports VLSM
• uses the Dijkstra shortest path algorithm - SPF
• collaborates with MPLS (Multi Protocol Label Switching) - maintains routing tables
• large OSPF networks should be divided into areas - reduces SPF computation, smaller routing tables, reduced LSU (link-state update)
• thanks to the areas, it supports a hierarchical network design
• uses a link-state database, which prevents loops
• manual summarization can be used to reduce routing tables, but automatic summarization is not supported
• routing updates are sent only when needed
• uses multicast (so it doesn't affect stations and other devices without OSPF), 224.0.0.6 - all DR and BDR, 224.0.0.5 - all routers, 224.0.0.2 - all routers on the same subnet
• does not use TCP, but IP protocol 89, it handles error detection and correction itself
• supports authentication, plain text or MD5, for accepting LSA
• uses a hello protocol - better than exchanging tables in RIP, uses it to establish adjacency, sends every 10s
• for routers to become neighbors, the hello packet must have matching values - area number, area type (stub or NSSA flag), subnet and subnet mask, hello and dead timer, authentication data
• LSA is sent every 30min or upon change, spread using flooding
• link cost for OSPF cost = 100Mb / bandwidth
• OSPF router ID (set at restart) is the highest IP address (active) on the router or the loopback interface address (this is recommended, as it is always up)
• DR/BDR election is based on the configured priority or Router ID, occurs only in multi-access networks, not in point-to-point, when the DR fails, the BDR becomes the DR and a new BDR is elected, if the original DR comes back online, it does not automatically become the DR (only when the DR and BDR both fail)
• if an area cannot be directly connected to the backbone, it can be connected to another area that is directly connected to area 0, and a virtual link must be created (a logical path to the backbone through the transit area) - identified by the neighbor router's router ID (if set incorrectly, it won't be created)
• area boundaries are defined within the router
• up to 30 OSPF routing processes can run on a router

Tables for OSPF

• routing - the best routes to destinations
• topology - routing records for all destinations
• neighbors - information about neighboring routers
• link-state database - the same for all routers, synchronized through LSA flooding, used to create routing tables, a tree-structured image of the network topology

OSPF Routes (Record Types)

• O - OSPF network within the same area (intra-area)
• O IA - OSPF inter-area, from ABR (LSA 3 and 4)
• O* IA - default OSPF inter-area
• O N1 - OSPF NSSA external type 1
• O* N2 - default OSPF NSSA external type 2
• O E1 - OSPF external type 1, from ASBR,
• O* E2 - default OSPF external type 2

OSPF Area Types

• standard area - accepts link updates, summary routes, and external routes
• backbone area - connects all other areas, always Area 0, same properties as standard
• stub area - does not receive routes from other AS (autonomous system), a default route is used for routing outside the AS, O, O IA, O* IA
• totally stubby area - Cisco proprietary, does not receive summary routes outside its area, O, O* IA
• not-so-stubby area (NSSA) - similar to stub, but imports some external type 7 LSA routes and converts them to type 5 LSA, O, O IA, O N2, O* N2

OSPF Router Types

• Area Border Router - ABR - has interfaces in multiple areas, has a separate link-state table for each area, connects areas to the backbone
• Autonomous System Border Router - ASBR - has interfaces in multiple AS, used to distribute routes from another AS, often also runs BGP
• Internal Router - regular, only in one area
• Backbone Router - at least one interface in Area 0

DR/BDR Routers

• Designated Router - DR - sends LSA 2 to all neighbors in multi-access, it is the router (or more precisely, the router interface) that is elected within the segment in multi-access (using special techniques even in NBMA), serves to reduce network traffic, the DR is the source of routing updates, maintains a complete topology table, all other routers establish a connection with it
• Backup Designated Router - BDR - becomes the DR if the original DR fails, has the second highest priority during the election

Link State Advertisements - LSA in OSPF

• type 1 - router, information about the router and its directly connected interfaces, only within the area, generated by each router
• type 2 - network, information about the LAN and the routers in it, in a multi-access network it originates from the DR, only within the area
• type 3 - summary, originates from the ABR (Area Border Router), networks accessible outside the area, for the ABR
• type 4 - ASBR summary, originates from the ABR, for the ASBR
• type 5 - external AS, announces external routes (default route), originates from ASBR, within the AS
• type 6 - multicast info
• type 7 - other extensions - NSSA

Network Types

• broadcast multi-access (BMA) - shared medium, Ethernet, DR and BDR are elected automatically
• nonbroadcast multi-access (NBMA) - without subinterfaces, typically serial Frame Relay, DR and BDR require manually defined neighbors
• point-to-point (PtoP)- connection between two routers, does not require DR and BDR
• point-to-multipoint (PtoM) - one interface that connects to multiple destinations (using subinterfaces), a series of point-to-point circuits, no DR/BDR election, often serial Frame Relay, automatically forms adjacency (no need for manual definition)

Cisco IOS Commands for Configuring OSPF

ROUTER(config)#router ospf 1         // 1 is the process-id, we can run multiple processes on a router

ROUTER(config-router)#network 192.168.5.4 0.0.0.3 area 1 // uses wildcard mask, 1 is the area number
ROUTER(config-router)#neighbor 192.168.5.4     // defines a neighbor, can include priority/cost
ROUTER(config-router)#summary-address 10.1.0.0 255.255.0.0 // summarize networks on ASBR
ROUTER(config-router)#redistribute connected [subnets]  // sends all local interfaces, subnets - also sends subneted routes
ROUTER(config-router)#passive-interface Serial0/0       // the given interface does not send or receive updates

ROUTER(config-router)#area 1 stub              // designates area 1 as a stub
ROUTER(config-router)#area 1 stub no-summary   // does not send LSA 3 and 4 = totally stubby 
ROUTER(config-router)#area 1 nssa no-summary   // does not send LSA 3 and 4 = NSSA totally stubby 
ROUTER(config-router)#area 1 range 10.1.0.0 255.255.0.0 // summarize networks on ABR, can specify whether to advertise - keyword advertise
ROUTER(config-router)#area 16 virtual-link 8.187.175.82 // virtual link to IP (e.g. loopback) into the backbone

ROUTER(config-if)#ip ospf network point-to-multipoint   // sets the mode on the interface, other options are broadcast, non-broadcast, point-to-point
ROUTER(config-if)#encapsulation frame-relay             // sets the encapsulation on the interface
ROUTER(config-if)#ip ospf priority 10                   // sets the priority for DR election
ROUTER(config-if)#ip ospf cost 10        // sets the cost of outgoing packets on the interface
ROUTER(config-if)#frame-relay map ip 10.1.1.1 200 broadcast // 200 = DLCI (Data-link connection identifier), broadcast is needed for Frame Relay

ROUTER#show ip ospf                // main information about the OSPF process, areas, etc.
ROUTER#show ip ospf border-routers // internal routing records to ABR and ASBR
ROUTER#show ip ospf neighbor       // information about neighbors per interface including link state, excluding ABR, ASBR, SPF
ROUTER#show ip ospf interface      // information from the given interface related to OSPF (link state, etc.) - router ID, neighbor relationship
ROUTER#show ip ospf virtual-link   // info about the virtual link to the backbone
ROUTER#show ip ospf database       // info about topology, link state, LSA

OSPF and IPv6

• IPv6 is supported in OSPF version 3 (and only for IPv6, not IPv4), the basis is similar to the previous OSPF versions, but there is a new configuration mode
• we can run multiple OSPF IPv6 instances
• the network command is not used to include an interface in the routing process, but the interface is configured directly

ROUTER#show ipv6 ospf                // main information about the OSPF process, areas, etc.
ROUTER#clear ipv6 ospf process       // the OSPF DB will be refreshed and the SPF will be restarted
ROUTER(config)#ipv6 unicast-routing  // enables IPv6 support
ROUTER(config)#ipv6 router ospf 1    // enables OSPF IPv6, 1 - internal process-id identification 1 to 65535 
ROUTER(config-if)#ipv6 ospf 1 area 5 // enables OSPF IPv6 for the interface, 1 - process-id, 5 - area-id 
ROUTER(config-if)#ipv6 ospf authentication ipsec spi 500 md5 // IPsec authentication