Quick Guide - Ip Routing
- Typically, in a large network, a combination of both static and dynamic routing is used.
- A packet can also be dropped on the RETURN trip. Example -- if you ping a host and it
doesn't answer, that doesn't necessarily mean the forward path is broken.
- Static Routing -- an entry is added to the routing table of each router, for each remote
- Static Routing Advantages -- no overhead on the router CPU, no bandwidth usage between
routers, security. Primarily used in small networks (< id="fullpost">
- Default Routing -- can only be used on stub networks (only one exit port out of the
- Before setting up a default route, remove any static routing entries -- "no ip route
172.16.20.0 255.255.255.0 172.16.10.2".
- Set up a default route -- "ip route 0.0.0.0 0.0.0.0 172.16.10.2". Verify the changes -- "sh ip
route" -- the default route shows as "S*".
- ALWAYS remember to use "ip subnet-zero" and "ip classless" -- both are enabled by
default in Cisco IOS 12.x.
- Dynamic Routing -- happens automatically, but uses CPU and network resources.
- Dynamic Routing Protocols within a network:
- RIP -- Routing Information Protocol.
- IGRP -- Interior Gateway Routing Protocol -- Cisco proprietary.
- EIGRP -- Enhanced Interior Gateway Routing Protocol -- Cisco proprietary.
- OSPF -- Open Shortest Path First -- non-proprietary.
- Dynamic Routing Protocols across networks:
- IGP -- Interior Gateway Protocol -- routing between routers in the same Autonomous
- EGP -- Exterior Gateway Protocol -- routing between different Autonomous Systems. BGP
(Border Gateway Protocol) is an example of an EGP.
- Administrative Distance -- a number from 0 to 255, where 0 is most reliable, 255 is blocked.
- If two advertised routes for the same network have the same ADs, the router makes a
decision by looking at hop count or bandwidth.
- AD values:
- 0 -- Connected interface.
- 1 -- Static route.
- 90 -- EIGRP.
- 100 -- IGRP.
- 110 -- OSPF.
- 120 -- RIP.
- 170 -- External EIGRP.
- 255 -- Unknown -- this one will never be used.
- Routing Protocol Types:
- Distance Vector -- distance is measured in hops. Examples -- RIP, IGRP.
- Link State -- also called Shortest Path First -- 3 tables per router -- 1 for directly attached
neighbors, 1 for network topology, and 1 for routing. Examples -- OSPF.
- Hybrid -- these use a combination of both methods -- EIGRP.
- Distance Vector Routing Protocols -- they use "routing by rumor" -- exchange of routing
tables. RIP looks at ADs first, then at hop count. If everything is equal, it performs round-
robin load balancing for up to 6 equal cost links.
- Pinhole Congestion -- with RIP routing, if a 56K link has less hops than a T1 link, the 56K
link will be used -- this is bad, and happens because hop count is the only metric used with
- Slow Convergence is another problem of RIP. While the routers are converging
(synchronizing their routing tables), no data is passed.
- RIP Routing table fields -- network number, exit interface, and hop count.
- Routing loops are also a common problem in RIP. They result from the slow convergence of
- Maximum Hop Count -- RIP has this set to 15. After that, a packet is dropped. Maximum
Hop Count is a good feature to decrease the severe effects of routing loops.
- Split Horizon -- information cannot be sent back in the direction from which it was received.
This method is able to prevent routing loops.
- Route Poisoning -- when a network becomes unreachable, the first directly attached router
places a "16" entry (unreachable) for this network in its routing table, and then advertises it to
all other routers. They reply with a "poison reverse" (acknowledgement).
- Holddowns -- these prevent regular update messages from a flapping network, router, or
interface. Thus, the flow of information continues.
- Holddown behavior:
- Holddowns have a timer. When it expires, the link is reinstated.
- If another update is received, with a better metric, the link is reinstated. If the metric is the
same, nothing happens.
- If a flush timer removes the bad route from the routing table (if it happens to expire --
coincidence), the link is reinstated.
- RIP is a true distance-vector routing protocol. It sends the complete routing table to all
active interfaces every 30 seconds.
- RIP Version 1 uses only classful routing. RIP Version 2 provides prefix routing (classless
routing) -- no subnet mask is sent with the updates.
- RIP Timers:
- Route update timer -- how often to send out updates -- default is 30 seconds.
- Route invalid timer -- when there are no updates for a specific route over a time period
(default is 90 seconds), the route is advertised as invalid.
- Route flush timer -- how long after a route becomes invalid before it is removed from the
routing table -- default is 240 seconds.
- Configure RIP routing -- Make sure there are no static routes, as they take precedence. Then
-- "config t", "router rip", "network 172.16.0.0" -- "network" tells the router which network to
- RIP is configured with classful routing network addresses -- ALL subnet masks must be the
same on all devices on the network.
- "sh ip route" displays something like this -- "R 172.16.50.0 [120/3] via 172.16.10.2,
FastEthernet0/0" -- "[120/3]" is the AD and the hop count.
- Blocking RIP advertisements after a certain point of the network -- "config t", "router rip",
"network 172.16.0.0", "passive-interface serial 0" -- serial 0 will stop advertising, but will still
- IGRP -- Cisco proprietary, maximum hop count of 255 with default of 100, helpful in larger
networks. IGRP uses bandwidth and delay of the line as metrics -- this combination is called a
- IGRP can also use other metrics, but they are not used by default -- reliability, load, and
- IGRP Timers:
- Update timer -- how frequently routing-update messages should be sent -- default is 90
- Invalid timer -- how long a router should wait before declaring a route invalid -- default is 3
x update timer.
- Holddown timer -- specifies the holddown period -- default is 3 x update timer + 10 seconds.
- Flush timer -- how long before a route is flushed from the routing table -- default is 7 x
- Configure IGRP -- "config t", "router igrp 10", "network 172.16.0.0" -- "10" is the
Autonomous System (AS). All routers must be in the same AS in order to communicate.
- You must ALWAYS use a classful network number when configuring IGRP. Example -- if
you type "172.16.10.0", the router will change it to "172.16.0.0". Still, DO NOT type
anything like this.
- IGRP can load balance up to 6 unequal links (while with RIP, they must be equal). The
"variance" command controls the load balancing between the best and the worst metric.
- If both RIP and IGRP are enabled on a router, it will always use IGRP, as IGRP has higher
precedence. Therefore, when using IGRP, disable RIP in order to spare resources.
- Commands to troubleshoot routing:
- "show ip route" -- displays the routing table.
- "show protocols" -- displays hardware information and link status.
- "show ip protocols" -- lots of routing information, including various parameters.
- "debug ip rip" -- sends debugging messages to the console. Can be redirected to the terminal
via "terminal monitor". Disable with "undebug all".
- "debug ip igrp events" -- debug summary of IGRP. Disable with "undebug" or "undebug
- "debug ip igrp transactions" -- full debug of IGRP. Again, disable with "undebug all".