Dynamic Route Configuration


Dynamic routing, in contrast to static routing, is where routing information is exchanged with other routers on the network. Entries in the routing table can change dynamically depending on the current status of the network. The protocols that are bundled with the Solaris 9 operating environment are Routing Information Protocol (RIP) versions 1 and 2 as well as the Router Discovery Protocol (RDISC).

Routing Information Protocol Version 1 (RIPv1)

RIP is a distance vector protocol , where the best route to use is calculated by the fewest hops taken to reach the destination. Each hop denotes a point where another router is used to forward a packet. The hopcount refers to the total number of hops taken to reach the final destination.

graphics/note_icon.gif

The number of hops required to reach a destination does not reflect the physical distance between two systems or networks; it merely reflects the number of routers that need to be used to deliver the data ”the fewer hops required, the less a route costs.


A number of routes might be available to use between two hosts , but RIP will keep a record of only the one that takes the fewest hops to reach. The dynamic nature of this protocol means that if an intermediate router becomes unavailable, then the best route can be modified to use the next best route.

RIP employs a number of features that help it cope with quickly changing networks:

  • Maximum hopcount ” RIP has a limit of 15 hops to prevent data traveling around the network endlessly. As a data packet traverses the network, the hopcount is incremented each time it passes through a router. If this value reaches 16 or more, the host is deemed unreachable.

  • Hold down timers ” When a router receives information about a failed router, a hold down timer starts. This prevents the receipt of further updates, which could result in an inconsistent network state. The time delay allows the network to stabilize with the new information before it calculates a new best route. It also allows the failed router time to recover. If a router has not yet received the failed router update, it continues to advertise the router as operational; the hold down timer prevents these invalid updates from being acted upon.

  • Split horizons ” This feature prevents a router from sending an update back to the same place it just came from. For example, if router A sends an update to router B, router B is prevented from advertising the information back to router A. Without this feature, two routers could become involved in an endless loop of information exchange.

  • Update triggers ” When a change to the network occurs, either because a new route has been found or an existing route has become unavailable, a router advertises this information immediately instead of waiting for the next scheduled advertisement ”by default every 30 seconds.

  • Route poisoning ” When a route becomes unavailable, an update is triggered by the router to advertise the route with a hopcount of 16. This advertises the route as unreachable, making the other routers remove the route entry from the route table. In this way the failed route will not be used.

The Routing Daemon

The daemon process that manages routing in Solaris 9 is called in.routed . It is started at boot time by the startup script /etc/rc2.d/S69inet (which is a hard link to /etc/init.d/inetinit ). Both routers and non-routers can run the in.routed process: Routers advertise their routing table entries to other hosts every 30 seconds, whereas non-routers learn the available routes by accessing the advertisements provided by other routers on the network.

graphics/alert_icon.gif

The time interval of 30 seconds between router advertisements is fixed and cannot be configured by the administrator.


The following are the most popular options for running this daemon:

  • -q ” Runs the daemon in "quiet" mode. Non-routers use this option to prevent the routing table from being advertised.

  • -s ” Runs the daemon in "speaking" mode, so that the routing table is broadcast (advertised) to other hosts every 30 seconds. This is the option a router uses.

  • -t ” Used in conjunction with the -q or “s options, this option causes all actions to be logged to the standard output. The following command starts the daemon process on a router and logs all activity to standard output.

     ultra10# /usr/sbin/in.routed -s -t 
  • -v <logfile> ” Used to log (with a timestamp) changes made to the routing table. The < logfile > location must also be entered. The following command starts the daemon process on a router and logs all changes to the file /var/log/route.log :

     ultra10# /usr/sbin/in.routed -v /var/log/route.log 
graphics/note_icon.gif

When running the routing daemon, in.routed , with the -v option, the logfile that is specified must be created. The process does not automatically create it.


graphics/alert_icon.gif

The in.routed process is started at boot time only if the files /etc/defaultrouter and /etc/notrouter are not present. The /etc/notrouter file is used to prevent a system with more than one network interface card from being treated as a router automatically.


Router Discover Protocol (RDISC)

The RDISC protocol is used to provide default routes between a host and the routers on the network. The daemon process that is run to discover routers is in.rdisc . In a similar way to in.routed , in.rdisc has two forms of execution: one for routers and one for non-routers. A router advertises its presence on the network every 10 minutes (600 seconds) by default using the local multicast address ( 224.0.0.1 ), although this value can be changed. A non-router listens to the advertisements and populates its routing table with the default route entries for each router. The in.rdisc process is started from the startup script /etc/rc2.d/S69inet , only if the following conditions are satisfied:

  • The system has multiple network interfaces configured.

  • The files /etc/defaultrouter and /etc/notrouter are not present.

  • The network interfaces are not being managed by the Dynamic Host Configuration Protocol (DHCP).

There are several advantages in using RDISC:

  • Resilience can be added to the network because of the existence of multiple default routes.

  • The default routes remove the need for multiple, often unnecessary, route entries, creating a smaller route table.

  • RDISC is not tied to any particular routing protocol, such as RIP; it is independent.

  • RDISC uses a multicast address rather than a broadcast address, resulting in less network traffic.

graphics/alert_icon.gif

Routers running the RDISC protocol still need to run a routing protocol, such as RIP. RDISC only provides a default route from a host to a router; it does not advertise between routers as a routing protocol does.


A non-router runs the in.rdisc process with the -s option to request information from routers on the network, as shown here:

 ultra10# /usr/sbin/in.rdisc -s 

A router uses the -r option to the in.rdisc command to cause it to advertise its presence on the network every 600 seconds by default. You can modify the interval by using the -T option. To start in.rdisc as a router and change the interval to 60 seconds (every minute), enter the following command:

 ultra10# /usr/sbin/in.rdisc -r -T 60 


Solaris 9 Network Administration Exam Cram 2 (Exam Cram CX-310-044)
Solaris 9 Network Administrator Exam Cram 2 (Exam CX-310-044)
ISBN: 0789728702
EAN: 2147483647
Year: 2003
Pages: 174
Authors: John Philcox

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