Routing
| Computer networks need to utilize some means of routing (or forwarding) mechanism so that communication can take place between two hosts , even though they may not be directly connected. Returning to the road analogy, routing loosely represents the use of maps. Take the example of traveling from Los Angeles, California, to a specific address in Las Vegas, Nevada. First, a map is needed to find out how to get to Las Vegas. Then, as the destination approaches, a more detailed map provides specific street information. Routing is similar in that, initially, the goal is to get to the corporate router on the required site. Other routers that are internal to the site direct the network packet to the correct network segment, and the packet arrives at the destination. Companies usually have one router that acts as an external interface to the rest of the world through which all incoming and outgoing traffic will pass. The external router also must be capable of routing traffic internally within the company to deliver packets destined for specific hosts. Internal routing can be managed in two ways: static routing, in which packets are sent to a central point for distribution; or dynamic routing, in which the best route is determined at the time the packet is sent. Both of these routing methods are discussed next . Static RoutingStatic routing involves each host on the network sending all network traffic to a specified address for further distribution. This address is entered into the file /etc/defaultrouter. When the system is restarted, the address found in this file will be used as the default "route" for all network traffic. The network administrator must manually build routing tables on the router so that data can be delivered anywhere on the local network. Static routing is used primarily on networks on which there is little or no change because it requires significant manual intervention when changes occur. For example, if the router address is changed, each host must have the file /etc/defaultrouter amended, and the network administrator must rebuild the routing tables to reflect the change. A disadvantage of static routing is that it creates a single point of failure ”that is, if the router fails, the network becomes unavailable because all the hosts have the address hard-coded in the /etc/defaultrouter file. Static routing, however, can be considered more secure than dynamic routing. This is because the default route address is fixed so that all traffic follows a predefined route; the traffic cannot be deviated by an unauthorized system being attached to the network, as is possible using dynamic routing where routes are advertised on the network. Dynamic RoutingIn an environment in which the structure of the network changes frequently, dynamic routing provides the flexibility to adapt to the changes with little or no manual intervention. Each host on the network runs a program that builds and maintains its own routing table. In Solaris, the program is the routing daemon ”in.routed for internal routing, or the gateway daemon, in.gated, if both internal and external routing are used. The route to use for network traffic is determined at the time the request is sent and is made more flexible by the fact that the host will "ask" its network neighbor for a route if it doesn't "know" how to send to the required destination. However, dynamic routing is not recommended for leaf nodes of a network ”that is, hosts not providing routing services for other systems. These nodes should use static routing so that the routers are not flooded with broadcasts from nonrouting hosts wanting to advertise their IP address. Dynamic routing is considerably more processor- intensive than static routing. However, it also requires much less manual intervention and is inherently more reliable than static routing because there is no single point of failure (as long as more than one route can physically be used). For example, if a major router fails and becomes unavailable, the routing algorithm will try to determine an alternate route to use (unlike static routing, in which the route is fixed). A potential disadvantage of dynamic routing is the advertising of routing information, a definite negative aspect for secure environments. It is possible, for example, for an unauthorized intruder to attach a host to the network and then to start acting as a router, advertising routes to destinations that don't exist, or capturing network traffic for malicious purposes. The dynamic nature of this routing method means that hosts might "learn" new routes from this bogus host. Static routing does not allow this to happen because of the fixed configuration of the central router. Several dynamic routing protocols exist, generally split into two categories:
|
| Top |
EAN: 2147483647
Pages: 101