OSPF

15-Oct-13

Topics

Background and features of OSPF Configure basic OSPF OSPF metric Designated router/backup designated router elections Default information originate

15-Oct-13

Routing protocols
Interior Exterior

Distance vector
RIP v1 RIP v2 IGRP EIGRP
15-Oct-13

Link state
OSPF IS-IS EGP BGP

OSPF background

Developed by IETF to replace RIP Better metric Fast convergence Scales to large networks by using areas

15-Oct-13

OSPF packets

0x01 Hello establishes and maintains adjacency 0x02 Database Description (DBD) summary of database for other routers to check 0x03 Link State Request (LSR) use to request more detailed information 0x04 Link State Update (LSU) reply to LSR and send new information 0x05 Link State Acknowledgement (LSAck)
15-Oct-13 5

OSPF encapsulation
Data link IP packet frame header header OSPF packet Data header

MAC destination address Multicast 01-00-5E-00-00-05 or 01-00-5E-00-00-06

15-Oct-13

OSPF encapsulation
Data link IP packet frame header header OSPF packet Data header

IP destination address Multicast 224.0.0.5 or 224.0.0.6 Protocol field 89

15-Oct-13

OSPF encapsulation
Data link IP packet frame header header OSPF packet Data header

Type code for packet type (0x01 etc) Router ID and Area ID

15-Oct-13

Hello, OSPF packet type 1

Discover OSPF neighbours and establish adjacencies. Advertise parameters on which two routers must agree to become neighbors. Elect the Designated Router (DR) and Backup Designated Router (BDR) on multiaccess networks like Ethernet and Frame Relay.
15-Oct-13 9

Fields in Hello packet

Type (=1), Router ID, Area ID Subnet mask of sending interface Hello Interval, Dead Interval Router Priority: Used in DR/BDR election Designated Router (DR): Router ID of the DR, if any Backup Designated Router (BDR): Router ID of the BDR, if any List of Neighbors: lists the OSPF Router ID of the neighboring router(s)
15-Oct-13 10

Sending Hellos

By default, OSPF Hello packets are sent every 10 seconds on multiaccess and pointto-point segments and every 30 seconds on non-broadcast multiaccess (NBMA) segments (Frame Relay, X.25, ATM). In most cases, OSPF Hello packets are sent as multicast to 224.0.0.5. Router waits for Dead interval before declaring the neighbor "down." Default is four times the Hello interval.
15-Oct-13 11

Matching

Before two routers can form an OSPF neighbour adjacency, they must agree on three values: Hello interval, Dead interval, Network type (e.g. point to point, Ethernet, NBMA.)

15-Oct-13

12

Election

On multi-access networks (Ethernet, NBMA) the routers elect a designated router and a backup designated router This saves on overhead Each router becomes adjacent to the designated router and swaps updates with it If the designated router fails, the backup designated router takes over
15-Oct-13 13

Finding best routes

15-Oct-13

14

Administrative Distance

Preferred to IS-IS or RIP but not to EIGRP

15-Oct-13

15

Configuring OSPF

R1(config)#router ospf 1 R1(config-router)# The process-id is between 1 and 65535 It does not have to match the process-id on neighbour routers (unlike EIGRP)

15-Oct-13

16

Configuring OSPF

Router(config-router)#network 192.168.1.0 0.0.0.255 area 0 Address as usual Wildcard mask is required (optional for EIGRP), some routers accept subnet mask We always use a single area 0 for CCNA, this would be the backbone if there are multiple areas.
15-Oct-13 17

Choosing the Router ID

2.

3.

Use the IP address configured with the OSPF router-id command. If the router-id is not configured, use the highest IP address of any of the loopback interfaces. If no loopback interfaces are configured, use the highest active IP address of any physical interface. The interface must be up. It need not be in a network command.
15-Oct-13 18

Show the router ID

show ip protocols (on most routers). show ip ospf show ip ospf interface

15-Oct-13

19

Loopback address

Highest loopback address is used in preference to a real interface address A loopback address is a virtual interface and is automatically up, so it cannot fail this makes it more stable. Router(config)#interface loopback 0 Router(config-if)#ip address 10.0.0.1 255.255.255.255
15-Oct-13 20

OSPF router-id command

Introduced in IOS 12.0(T) and is the first choice for determining router ID. Router(config)#router ospf 1 Router(config-router)#router-id 172.16.0.1 Many networks still use the loopback address method of assigning router IDs.
15-Oct-13 21

Changing router ID

The router ID is fixed when OSPF is configured and given its first network command. Any loopback addresses or router-id commands should be given before configuring OSPF. Router#clear ip ospf process can be used, set the ID, then configure OSPF again. The router may need to be reloaded
15-Oct-13 22

Show ip ospf neighbor

Dead Time

Address

Interface

1
1

FULL/
FULL/

00:00:30 192.168.10.6 Serial0/1

OSPF priority

Of neighbour Fully adjacent On this router

15-Oct-13

Other show commands

show ip protocols show ip ospf show ip ospf interface Show ip route

15-Oct-13

24

Summary?

OSPF does not summarise to class boundaries by default.

15-Oct-13

25

OSPF metric

The OSPF specification says that cost is the metric, does not say how cost is found. Cisco uses bandwidth Cost = 108 = 100,000,000 bandwidth bandwidth Then finds cumulative cost for all links on a path.

15-Oct-13

26

Standard costs
Interface type Fast Ethernet and faster Ethernet E1 T1 128 Kbps 64 Kbps 56 Kbps 108/bps = Cost 108/100,000,000bps = 1 108/10,000,000bps = 10 108/2,048,000bps 108/1,544,000bps 108/128,000bps 108/64,000bps 108/56,000bps = 48 = 64 = 781 = 1562 = 1785

15-Oct-13

27

Faster than 100 Mbps

By default, the cost metric for all interfaces operating at 100Mbps or more is 1. This uses the reference bandwidth of 100Mbps. To distinguish between links of higher bandwidths, configure all routers in the area e.g. auto-cost reference-bandwidth 1000 This would multiply costs by 10 and allow for faster bandwidths to have costs below 10.
15-Oct-13 28

Serial link bandwidths

Serial links often have a default bandwidth of T1 (1.544 Mbps), but it could be 128 kbps. This may not be the actual bandwidth. show interface will give the default value. show ip ospf interface gives the calculated cost. Give it the right bandwidth. Router(config-if)#bandwidth 64
15-Oct-13 29

Configure the cost directly

Alternative to configuring the bandwidth: Configure the cost directly. R1(config)#interface serial 0/0 R1(config-if)#ip ospf cost 1562 Configure cost if there are non-Cisco routers in the area that calculate costs in different ways.

15-Oct-13

30

Point to point network

Only two routers on network They become fully adjacent with each other

15-Oct-13

31

Multiaccess networks

Networks where there could possibly be more than 2 routers, e.g. Ethernet, Frame Relay. These have a method of cutting down on adjacencies and the number of updates exchanged. 5 routers: 10 adjacencies?

15-Oct-13

32

Multiaccess network

Not efficient if they every router becomes fully adjacent to every other router Designated router (DR) becomes fully adjacent to all other routers Backup designated router (BDR) does too in case designated router fails

15-Oct-13

33

Multiaccess

All routers send LSUs to DR and BDR but not to other routers Use multicast address 224.0.0.6

DROther
15-Oct-13

DROther

DROther
34

Multiaccess

DR then sends LSUs to all routers Use multicast address 224.0.0.5

15-Oct-13

35

Router detects change

A router knows that a link is down if it does not receive a timed Hello from a partner

15-Oct-13

36

Send update

The router sends a LSU (link state update) on multicast 224.0.0.6 to DR/BDR

15-Oct-13

37

Update all routers

DR sends to 224.0.0.5, all OSPF routers BDR does not send unless DR fails

15-Oct-13

38

Recalculate routing table

Each router sends LSAck acknowledgement Waits for hold time in case link comes straight back up Runs SPF algorithm using new data Updates routing table with new routes

15-Oct-13

39

OSPF network types

Characteristics Ethernet, token ring, FDDI Frame relay, X.25, ATM PPP, HDLC Configured by administrator Configured by administrator

DR election? Yes

Yes
No No No
40

DR/BDR election

Happens when routers first discover each other using Hellos. Router with highest priority becomes DR, next highest becomes BDR. If they have the same priority then the highest router ID becomes DR, next highest becomes BDR. By default all routers have priority 1
15-Oct-13 41

Election where same priority

15-Oct-13

42

Add a router

An election has taken place and a DR and BDR have been chosen. Now add another router with a higher priority. It will not become DR if there is already a DR. To make sure that a certain router becomes DR:

Give it the highest priority Switch it on first

15-Oct-13

OSPF states

Down Init (after receiving hello) Two-way (election here) ExStart (decide who initiates exchange) Exchange (swap summary database) Loading (link state requests and updates) Full adjacency (know the same topology)
15-Oct-13 44

DROther routers

Routers that are not elected as DR or BDR are called DROther. They become fully adjacent with DR and BDR. They stay in 2-way state with each other.

15-Oct-13

45

Priority

Router(config-if)#ip ospf priority {0 - 255} To force an election: Shut down the interfaces Bring them up again, chosen DR first, chosen BDR second. The DR should be a router with plenty of processing power.

15-Oct-13

46

Propagate static route

R1(config-router)#default-information originate

In routing table O*E2 0.0.0.0/0 [110/1] via 192.168.10.10, 00:05:34, Serial0/0/1 E2 means this is an OSPF External Type 2 route. The cost will stay the same as it is propagated. Type 1 would increase its cost at each router.

15-Oct-13

47

Changing intervals

Router(config-if)#ip ospf hellointerval seconds Router(config-if)#ip ospf deadinterval seconds This needs to be done on both partners in an adjacency. The adjacency is broken when one router is changed.
15-Oct-13 48

Databases

15-Oct-13

49

Comparing routing protocols

Distance vector

Sends LSA updates low bandwidth use after initial flooding Complex algorithm powerful processor Three databases large memory No loops
15-Oct-13

Broadcasts whole routing tables high bandwidth use Simple algorithms little processing One table little memory Can have loops
50