IGRP ( Interior Gateway Routing Protocol )

• The topological delay time


• The bandwidth of the narrowest bandwidth segment of the path


• The channel occupancy of the path


• The reliability of the path

• Upon router startup, IGRP broadcast request messages to other routers. Upon receipt of the message, router send their routing tables to the startup router. Routing tables contain entries for each of the networks that a router can reach.



• At regular intervals, routers automatically transmit their routing tables to other routers which in turn update their own routing tables if there are changes in the topology.



• A triggered update occurs if the network topology changes. Triggered updates allow quick responses to changes in the network.



• When routers add entries to routing tables, 1 is added to the hop count to account for the hop between the router and the neighbor from which it received the route information.



• When a route is entered into the routing table, a timer is set. As routing table updates arrive and a route entry is verified to still be valid, the timer is reset. If a known router fails to appear in an update and in subsequent updates, the timer will run out and the route entry will be purged.

• Holddowns

Holddowns are used to prevent regular update messages from inappropriately reinstating a route that might have gone bad. When a router goes down, neighboring routers detect this via the lack of regularly scheduled update messages. These routers then calculate new routes and send routing update messages to inform their neighbors of the route change. This activity begins a wave of triggered updates that filter through the network. These triggered updates do not instantly arrive at every network device. Thus, it is possible for a device that has yet to be informed of a network failure to send a regular update message, which advertises a failed route as being valid to a device that has just been notified of the network failure. In this case, the latter device would contain (and potentially advertise) incorrect routing information. Holddowns tell routers to hold down any changes that might affect routes for some period of time. The holddown period usually is calculated to be just greater than the period of time necessary to update the entire network with a routing change.


• Triggered updates

Triggered updates would be sufficient if we could guarantee that the wave of updates reached every appropriate gateway immediately. However, there are two problems. First, packets containing the update message can be dropped or corrupted by some link in the network. Second, the triggered updates don't happen instantaneously. It is possible that a gateway that has not yet gotten the triggered update will issue a regular update at just the wrong time, causing the bad route to be reinserted in a neighbor that had already gotten the triggered update. Holddowns are designed to get around these problems. The holddown rule says that when a route is removed, no new route will be accepted for the same destination for some period of time. This gives the triggered updates time to get to all other gateways, so that we can be sure any new routes we get aren't just some gateway reinserting the old one. The holddown period must be long enough to allow for the wave of triggered updates to go throughout the network. In addition, it should include a couple of regular broadcast cycles, to handle dropped packets. Consider what happens if one of the triggered updates is dropped or corrupted. The gateway that issued that update will issue another update at the next regular update. This will restart the wave of triggered updates at neighbors that missed the initial wave.


• Split horizons

Split horizons derive from the premise that it is never useful to send information about a route back in the direction from which it came. For example, Router 1 (R1) advertises that it has a route to Network A. There is no reason for Router 2 (R2) to include this route in its update back to R1 because R1 is closer to Network A. The split-horizon rule says that R2 should strike this route from any updates that it sends to R1. The split-horizon rule helps prevent routing loops. Consider, e.g. the case in which R1's interface to Network A goes down. Without split horizons, R2 continues to inform R1 that it can get to Network A (through R1). If R1 does not have sufficient intelligence, it actually might pick up R2's route as an alternative to its failed direct connection, causing a routing loop. Although holddowns should prevent this, split horizons are implemented in IGRP because they provide extra algorithm stability.

Split horizons should prevent routing loops between adjacent routers, but poison-reverse updates are necessary to defeat larger routing loops. Increases in routing metrics generally indicate routing loops. Poison-reverse updates then are sent to remove the route and place it in holddown. In Cisco's implementation of IGRP, poison-reverse updates are sent if a route metric has increased by a factor of 1.1 or greater.

• Version

Protocol version number (currently 1).


• Opcode

Operation code indicating the message type:

1 Update.
2 Request.


• Edition

Serial number which is incremented whenever there is a change in the routing table. The edition number allows gateways to avoid processing updates containing information that they have already seen.


• ASystem

Autonomous system number. A gateway can participate in more than one autonomous system where each system runs its own IGRP. For each autonomous system, there are completely separate routing tables. This field allows the gateway to select which set of routing tables to use.
Ninterior, Nsystem, Nexterior
Indicate the number of entries in each of these three sections of update messages. The first entries (Ninterior) are taken to be interior, the next entries (Nsystem) as being system, and the final entries (Nexterior) as exterior.


• Checksum

IP checksum which is computed using the same checksum algorithm as a UDP checksum. The checksum is computed on the IGRP header and any routing information that follows it. The checksum field is set to zero when computing the checksum. The checksum does not include the IP header and there is no virtual header as in UDP and TCP.
An IGRP request asks the recipient to send its routing table. The request message has only a header. Only the Version, Opcode and ASystem fields are used; all other fields are zero.
An IGRP update message contains a header, immediately followed by routing entries. As many routing entries as possible are included to fit into a 1500-byte datagram (including the IP header). With current structure declarations, this allows up to 104 entries. If more entries are needed, several update messages are sent.

• Stable routing even in very large or complex networks. No routing loops should occur, even as transients.


• Fast response to changes in network topology.


• Low overhead. That is, IGRP itself should not use more bandwidth than what is actually needed for its task.


• Splitting traffic among several parallel routes when they are of roughly equal desirability.


• Taking into account error rates and level of traffic on different paths.


• The ability to handle multiple "types of service" with a single set of information.

• Create the IGRP Routing Process

To create the IGRP routing process, perform the following required tasks starting in global configuration mode:


Command	Task
router igrp autonomous-system	Step 1 Enable an IGRP routing process, which places you in router configuration mode.
Network network-number	Step 2 Associate networks with an IGRP routing process.


IGRP sends updates to the interfaces in the specified networks. If an interface's network is not specified, it will not be advertised in any IGRP update.


• Allow Point-to-Point Updates for IGRP

Because IGRP is normally a broadcast protocol, in order for IGRP routing updates to reach point-to-point or nonbroadcast networks, you must configure the router to permit this exchange of routing information.

To permit information exchange, perform the following task in router configuration mode:


Command	Task
Nerighbor ip-address	Define a neighboring router with which to exchange point-to-point routing information.



• Define Unequal-Cost Load Balancing

IGRP can simultaneously use an asymmetric set of paths for a given destination. This feature is known as unequal-cost load balancing. Unequal-cost load balancing allows traffic to be distributed among multiple (up to four) unequal-cost paths to provide greater overall throughput and reliability. Alternate path variance (that is, the difference in desirability between the primary and alternate paths) is used to determine the feasibility of a potential route. An alternate route is feasible if the next router in the path is closer to the destination (has a lower metric value) than the current router and if the metric for the entire alternate path is within the variance. Only paths that are feasible can be used for load balancing and included in the routing table. These conditions limit the number of cases in which load balancing can occur, but ensure that the dynamics of the network will remain stable.

The following general rules apply to IGRP unequal-cost load balancing:

• IGRP will accept up to four paths for a given destination network.



• The local best metric must be greater than the metric learned from the next router; that is, the next-hop router must be closer (have a smaller metric value) to the destination than the local best metric.



• The alternative path metric must be within the specified variance of the local best metric. The multiplier times the local best metric for the destination must be greater than or equal to the metric through the next router.


• Control Traffic Distribution

IGRP or Enhanced IGRP will distribute traffic among multiple routes of unequal cost to the same destination. If you want to have faster convergence to alternate routes but you do not want to send traffic across inferior routes in the normal case, you might prefer to have no traffic flow along routes with higher metrics.

To control how traffic is distributed among multiple routes of unequal cost, perform the following task in router configuration mode:


Command	Task
Traffic-share {balanced | min}	Distribute traffic proportionately to the ratios of metrics, or by the minimum-cost route.



• Adjust the IGRP Metric Weights

You have the option of altering the default behavior of IGRP routing and metric computations. This allows, for example, tuning system behavior to allow for transmissions via satellite. Adjusting IGRP metric weights can dramatically affect network performance, however, so ensure you make all metric adjustments carefully.

To adjust the IGRP metric weights, perform the following task in router configuration mode. Due to the complexity of this task, we recommend that you only perform it with guidance from an experienced system designer.


Command	Task
metric weights tos k1 k2 k3 k4 k5	Adjust the IGRP metric.


By default, the IGRP composite metric is a 24-bit quantity that is a sum of the segment delays and the lowest segment bandwidth (scaled and inverted) for a given route.


• Enforce a Maximum Network Diameter

The router enforces a maximum diameter to the IGRP network. Routes whose hop counts exceed this diameter will not be advertised. The default maximum diameter is 100 hops. The maximum diameter is 255 hops.

To configure the maximum diameter, perform the following task in router configuration mode:


Command	Task
metric maximum-hops hops	Configure the maximum network diameter.



• Validate Source IP Addresses

To disable the default function that validates the source IP addresses of incoming routing updates, perform the following task in router configuration mode:


Command	Task
metric maximum-hops hops	Configure the maximum network diameter.

• Backward compatible with IGRP, allowing EIGRP to be introduced into existing IGRP networks.



• EIGRP provides fast convergence since routers store all neighbor routing tables to promote faster route updates if necessary.



• Supports variable-length subnet mask, allowing route summarization (aggregation) and the creation of network boundaries.



• Supports partial updates so that entire routing tables do not need to be sent at regular intervals. Updates are sent when a route changes and the changes are sent to only the routers that need the information.



• Supports multiple network-layer protocols, including AppleTalk, Novell NetWare, and IP. The IP implementation of EIGRP redistributes reoutes learned from OSPF, RIP, IS-IS, EGP, and BGP.



• EIGRP uses neighbor discovery/recovery to dynamically learn about other routers on directly attached networks. Hello packets are periodically sent to monitor the status of links. EIGRP also uses a reliable transport protocol to provide guaranteed, ordered delivery of packets to neighbors. Route computations are handled by DUAL, which uses distance information to select efficient, loop-free paths.

IP Implementation



This section gives a brief description of the packet formats used by Cisco's IGRP. IGRP is 

sent using IP datagrams with IP protocol 9 (IGP). The packet begins with a header. It 

starts immediately after the IP header. 



          unsigned version: 4;  /* protocol version number */

          unsigned opcode: 4;   /* opcode */

          uchar edition;        /* edition number */

          ushort asystem;       /* autonomous system number */

          ushort ninterior;     /* number of subnets in local net */

          ushort nsystem;       /* number of networks in AS */

          ushort nexterior;     /* number of networks outside AS */

          ushort checksum;      /* checksum of IGRP header and data */

■ Parent layer
■ Child layer

IP

IGRP