The same applies with protocols such as ping. Ping sends a echo to the destination, and when the destination device receives the ICMP packet it must send a echo-reply packet back to the source. If there is no route going back to the source, ping will be unsuccessful.

In the live demo below I have illustrated this exact problem. On my router AOIP.ORG I have tried to ping the IP address 2.2.2.2 (the loopback interface of R2). AOIP.ORG has a default gateway of last resort configured to send all traffic to the router R1 (R1 already has routing configured correctly), R1 will send the packet to R2 however R2 does not have a route back to the 192.168.1.0 network (attached to AOIP.ORG). Only after I telnet into R1, and then telnet into R2 and configure a route to the network 192.168.1.0 do my pings begin to work. ( I can’t telnet directly to R2 since it does not have a route back to me, so I have to do this on a hop by hop basis).

Related posts:

1. Static Routing
2. Inter-VLAN Routing (Router on a Stick)
3. Configuring EIGRP on a Cisco Router

OSPF is a classless routing protocol so subnetmask values are sent in the update and it supports CIDR.

OSPF does not send periodic updates and is designed to only send updates when something has changed or a new route has been added, it does this through the use of LSA’s (Link State Advertisements).

 OSPF is designed in a hierarchical fashion through Areas. Routers running OSPF, flood LSA updates to routers in the same area rather than sending the entire routing table.

 OSPF is a link-state protocol, and pieces the information received from LSA’s to calculate the best path using the SPF (Shortest Path First) algorithm. Its cost value (metric) is based on the formula: Cost = 10⁸/bandwidth (bps)

 Each router running the OSPF process has a Router ID which is known by all other devices. By default the Router ID is the highest IP address of an active interface at the point the OSPF process started. This can be manually forced by using loopback interfaces, as they are permanently in a Up Up status, and insuring the loopback has the highest IP address on the router. The default value can also be overwritten by using the router-id command

 OSPF has a default administration distance of 110

 OSPF does not do automatic summarization of routes like EIGRP does, but routes can be manually summarized as required.

In the Live demo below, the following configuration was entered.

 AOIP.ORG# ping 10.0.3.2

(ping is unsuccessful)

AOIP.ORG# show ip route

(no routes for network 10.0.3.0 exist)

AOIP.ORG# conf t

AOIP.ORG(config)# router ospf 100

(enables OSPF with a process ID of 100 – this value is used by THIS ROUTER ONLY)

AOIP.ORG(config-router)# network 192.168.1.0 0.0.0.255 area 0

(enables all interfaces with IP addresses beginning with 192.168.1 to belong to Area 0 – the backbone area. OSPF uses wildcard mask values for the network statement)

AOIP.ORG(config-router)# exit

AOIP.ORG(config)# exit

AOIP.ORG# ping 10.0.3.2 repeat 15

(ping is successful after a few attempts, it can take OSPF a little while to discover and learn the topology of the network)

AOIP.ORG# show ip route

(3 new entries are in the routing table with a ‘O’ to represent OSPF, the entries are not summarized by default, 10.0.1.0, 10.0.2.0 and 10.0.3.0)

Related posts:

1. Configuring EIGRP on a Cisco Router
2. Configuring RIP on a Cisco Router
3. Cisco Administrative Distance

EIGRP is a classless routing protocol so subnetmask values are sent in the update and it does support CIDR.

EIGRP does not send periodic updates and is designed to only send updates when something has changed or a new route has been added, it informs only routers that require the information about the change instead of flooding the entire network and only the change is sent to the routers not the entire routing table.

EIGRP is also able to load balance traffic over 6 different paths but it’s configured for 4 by default. It is also capable of doing unequal cost load balancing allowing it to send packets in proportional amounts to the bandwidth of the link. Unequal load balancing is not a default setting and needs to be configured using the Variance command.

Although EIGRP is often considered a distance vector protocol, it has a lot of the features of a link-state protocol and technically falls under a Hybrid protocol classification. It’s metric value is derived from 5 different aspects.

1. Bandwidth-Minimum bandwidth on the route, in kilobits per second.
2. Delay-Route delay, in tens of microseconds.
3. Reliability-Likelihood of successful packet transmission, expressed as an integer from 0 to 255.
4. Loading-Effective bandwidth of path.
5. MTU-Minimum MTU in path, expressed in bytes.

EIGRP takes each of the above values and runs them though the following formula to calculate the total metric value.

metric = [k1 x bandwidth + (k2 x bandwidth) / (256 - load) + k3 x delay]

If k5 does not equal 0 (default value), an additional calculation is done

metric = metric x (k5/(reliability + k4))

The default K values are k1 = k3 = 1 and k2 = k4 = k5 = 0.

NOTE: If the default values have not been changed the metric = bandwidth + delay.

EIGRP sends updates to routers that belong to the same A.S (Autonomous System) number, which can be looked at as a group number. If routers do not belong to the same group number (A.S) they will not share information with each other.

EIGRP by default does network summarization, this means that when the network crosses a network boundary the route will be summarized to its full class. In our network (Diagram below), we have a router named ‘AOIP.ORG’ which has an IP address of 192.168.1.10/24. The rest of the network is running on a 10.0.x.x/24 range, and EIGRP sees them as 2 network boundaries by default (192.168.1.0/24 and 10.0.0.0/8), when an update crosses either of those boundaries EIGRP will summarize the route into the full class. This default behaviour can be changed using the ‘no auto-summary’ command. The live demo below illustrates this issue.

In the live demo, I have done the following configuration

AOIP.ORG# ping 10.0.3.2                                                                              

(ping is unsuccessful)

AOIP.ORG# show ip route                                                                                  

(no route exists for the network 10.0.3.0)

AOIP.ORG# conf t                                                                                            

(enter global configuration)

AOIP.ORG(config)# router eigrp 100                                                       

(enable EIGRP with A.S number 100)

AOIP.ORG(config-router)# network 192.168.1.0 0.0.0.255             

(specifies which interface will send and receive updates, Wildcard mask is used by EIGRP. In this example any interface that has an IP address that starts with 192.168.1 will participate in EIGRP updates)

AOIP.ORG(config-router)# exit

AOIP.ORG(config)# exit

AOIP.ORG# show ip route                                                                                  

(The updated routing table now has a value of ‘D’ for the EIGRP entry, you will notice it is summarized as 10.0.0.0/8 as R1 is on the network boundary)

AOIP.ORG# ping 10.0.3.2                                                                              

(Confirmation that ping now works)

AOIP.ORG# telnet 192.168.1.1                                                                   

(telnet into R1 to disable Auto-Summary)

R1# conf t

R1(config)# router eigrp 100                                                                       

(Entering EIGRP AS 100 configuration)

R1(config-router)# no auto-summary                                                                         

(Disabling Auto-Summary)

R1(config-router)# exit

R1(config)# exit

R1# exit

AOIP.ORG# show ip route                                                                                  

(3 routes, 10.0.1.0, 10.0.2.0 and 10.0.3.0 now exist)

 

Related posts:

1. Configuring OSPF on a Cisco Router
2. Configuring RIP on a Cisco Router
3. Cisco Administrative Distance