Skip to main content

Load balancing across multiple paths

Load balancing across multiple paths
7.2.9 This page will further explain how routers use load balancing to transmit packets to a destination IP address over multiple paths. The paths are derived either statically or with dynamic protocols, such as RIP, EIGRP, OSPF, and IGRP.


When a router learns multiple routes to a specific network, the route with the lowest administrative distance is installed in the routing table. Sometimes the router must select a route from among many, learned through the same routing process with the same administrative distance. In this case, the router chooses the path with the lowest cost or metric to the destination. Each routing process calculates its cost differently and the costs may need to be manually configured in order to achieve load balancing.

If the router receives and installs multiple paths with the same administrative distance and cost to a destination, load-balancing can occur. Cisco IOS imposes a limit of up to six equal cost routes in a routing table, but some IGPs have their own limitations. EIGRP allows up to four equal cost routes.

By default, most IP routing protocols install a maximum of four parallel routes in a routing table. Static routes always install six routes. The exception is BGP, which by default allows only one path to a destination.

The range of maximum paths is one to six paths. To change the maximum number of parallel paths allowed, use the following command in router configuration mode:

Router(config-router)#maximum-paths [number ]

IGRP can load balance up to six unequal links. RIP networks must have the same hop count to load balance, whereas IGRP uses bandwidth to determine how to load balance.

In Figure , there are three ways to reach Network X:

• E to B to A with a metric of 30

• E to C to A with a metric of 20

• E to D to A with a metric of 45

Router E chooses the second path, E to C to A with a metric of 20, since it is a lower cost than 30 and 45.

Cisco IOS supports two methods of load balancing for IP packets. These are per-packet and per-destination load balancing. If process switching is enabled, the router will alternate paths on a per-packet basis. If fast switching is enabled, only one alternate route will be cached for the destination address. All packets that are bound for a specific host will take the same path. Packets bound for a different host on the same network may use an alternate route. Traffic is load balanced on a per-destination basis.

By default the router uses per-destination load balancing, also called fast switching. The route cache allows outgoing packets to be load-balanced on a per-destination basis rather than on a per-packet basis. To disable fast switching, use the no ip route-cache command. Using this command will cause traffic to be load balanced on a per-packet basis.

In the Lab Activities, students will configure and observe load balancing.

The next page will discuss static routes with RIP.

Comments

Popular posts from this blog

OSI layers / Peer-to-peer communications / TCP/IP model

OSI layers 2.3.4 This page discusses the seven layers of the OSI model. The OSI reference model is a framework that is used to understand how information travels throughout a network. The OSI reference model explains how packets travel through the various layers to another device on a network, even if the sender and destination have different types of network media. In the OSI reference model, there are seven numbered layers, each of which illustrates a particular network function. - Dividing the network into seven layers provides the following advantages: • It breaks network communication into smaller, more manageable parts. • It standardizes network components to allow multiple vendor development and support. • It allows different types of network hardware and software to communicate with each other. • It prevents changes in one layer from affecting other layers. • It divides network communication into smaller parts to make learning it easier to understand. In the foll...

PC Basic...

• Backplane – A backplane is an electronic circuit board containing circuitry and sockets into which additional electronic devices on other circuit boards or cards can be plugged; in a computer, generally synonymous with or part of the motherboard. • Network interface card (NIC) – An expansion board inserted into a computer so that the computer can be connected to a network. • Video card – A board that plugs into a PC to give it display capabilities. • Audio card – An expansion board that enables a computer to manipulate and output sounds. • Parallel port – An interface capable of transferring more than one bit simultaneously that is used to connect external devices such as printers. • Serial port – An interface that can be used for serial communication in which only one bit is transmitted at a time. • Mouse port – A port used to connect a mouse to a PC. • USB port – A Universal Serial Bus connector. A USB port connects devices such as a mouse or printer to the computer ...

1.2.2 RIP V2 Features

 1.2.2 RIP V2 Features This page will discuss RIP v2, which is an improved version of RIP v1. Both versions of RIP share the following features: It is a distance vector protocol that uses a hop count metric. It uses hold down timers to prevent routing loops – default is 180 seconds. It uses split horizon to prevent routing loops. It uses 16 hops as a metric for infinite distance. RIP v2 provides prefix routing, which allows it to send out subnet mask information with the route update. Therefore, RIP v2 supports the use of classless routing in which different subnets within the same network can use different subnet masks, as in VLSM. RIP v2 provides for authentication in its updates. A set of keys can be used on an interface as an authentication check. RIP v2 allows for a choice of the type of authentication to be used in RIP v2 packets. The choice can be either clear text or Message-Digest 5 (MD5) encryption. Clear text is the default. MD5 can be used t...