IGRP Metrics

IGRP is a distance-vector routing protocol designed for use within autonomous systems. IGRP, like other distance-vector routing protocols (such as RIP and RIPv2), forwards its entire routing table on a regular basis. IGRP, unlike other distance-vector protocols (RIP and RIPv2), uses a series of metrics rather than a single hop-count metric. These metrics are as follows:

  • Hop count

  • Packet size MTU (Maximum Transmission Unit)

  • Link bandwidth

  • Delay

  • Load

  • Reliability

IGRP supports multipath routing, load balancing across two, three, or four links, with automatic link recovery from a failed link.

Like RIP, IGRP is a classful routing protocol, with no support for VLSM (Variable Length Subnet Masking) or CIDR (Classless Inter-Domain Routing).

IGRP Hop Count

IGRP supports incrementing hop count as one method of determining the distance to a specific destination. Unlike RIP, which has a maximum hop count of 16, IGRP has a maximum hop count of 100, which may be further increased to 255.

IGRP does not use the hop counter as a metric to determine optimal paths to a destination, but rather as a means to detect routing loops in the network. Any route with a hop count in excess of the stated, or configured, maximum is automatically invalidated.

IGRP Packet Size (MTU)

The MTU (Maximum Transmission Unit) identifies the largest datagram that an IGRP router will accept for transmission. IGRP does not use the MTU to calculate a route or use it as a factor in its metrics. IGRP routers exchange MTU information with each other, identifying the maximum datagram size they can support. Datagrams larger than the MTU size will be broken down into manageable "pieces" by the router.

NOTE

Large MTU sizes may result in what is called a performance penalty. Routers will buffer these incoming datagrams for transmission until it can be determined where they are to be forwarded. The performance penalty lies in the delay to determine the next hop and the amount of router buffer (memory) space consumed waiting for the entire datagram to be received for transmission.

IGRP Link Bandwidth

IGRP uses the link bandwidth as one of the metrics to calculate the total cost of a network path. The default bandwidth statement is 1.544 Mbps (T1), but it can range from 1200 bps to 10 Gbps.

IGRP looks at the defined bandwidth on each outbound interface router port for a given route and will select the smallest bandwidth statement, which is the bandwidth limit for that route. This bandwidth number is then divided by 107 (10,000,000) to determine the bandwidth measurable in kilobits per second (kbps).

IGRP Delay

The IGRP delay metric measures the approximate amount of time needed to traverse a network link, based on no other link usage at the time. The aggregated delay metric of a route is the sum of all delays attributed to each outbound router interface in the path. This delay sum is divided by 10 to express the delay result in microseconds (µs). The metric value for delay ranges from 1 to 16,777,215 (224 1).

IGRP Load

The IGRP load metric measures the amount of bandwidth currently available across a given link. The heavier the link utilization, the more time is required for data traffic to traverse that link. This metric enables IGRP to factor current link utilization levels into the network's optimal route calculation.

This metric can be configured manually to any value between 1 and 255, although it is not recommended you do this, as there could be adverse impacts on network convergence and route distribution throughout the network.

IGRP Reliability

The reliability metric keeps track of the current error rate per transmission facility. The error rate is a ratio of packets received by a destination without error.

This metric is between the range of 1 and 255; the higher the metric value, the more unreliable the link.



Network Sales and Services Handbook
Network Sales and Services Handbook (Cisco Press Networking Technology)
ISBN: 1587050900
EAN: 2147483647
Year: 2005
Pages: 269

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net