Route maps can be used for both redistribution and for policy routing. They are also used frequently in large-scale Border Gateway Protocol (BGP) implementations . Although redistribution has been covered extensively in previous chapters, this chapter introduces the topic of policy routing. Policy routes are nothing more than sophisticated static routes. Whereas static routes forward a packet to a specified next hop based on the destination address of the packet, policy routes forward a packet to a specified next hop based on the source of the packet. Policy routes can also be linked to extended IP access lists so that routing may be based on such things as protocol types and port numbers . Like a static route, a policy route influences the routing only of the router on which it is configured. Note policy routing Figure 14.1 shows an example of a typical policy routing application. AbnerNet is connected to two Internet service providers via router Dogpatch. AbnerNet's corporate policy dictates that some users' Internet traffic should be sent via ISP 1 and other users' Internet traffic should be sent via ISP 2. If either ISP should become unavailable, the traffic normally using that provider will be sent to the other provider. A policy route at Dogpatch can distribute Internet traffic in accordance with local policy. The distribution of traffic might be based on subnet, specific user, or even user applications. Figure 14.1. Policy routing allows traffic from AbnerNet to be routed to one of its two Internet service providers based on parameters such as source address, source/destination address combinations, packet size , application-level ports, or even packet length.
Figure 14.2 shows another use for policy routing. One of the systems on the right watches for invasion forces from the planet Mongo while the other stores back copies of Dilbert comic strips . Policy routes can be configured to route the critical traffic from the Mongo System to Flash_G over the FDDI link and to route the lower-priority Dilbert traffic over the 56K links. Or vice versa. Figure 14.2. Policy routing allows high-priority traffic from the Mongo System to be routed over the FDDI link while low-priority traffic from the Dilbert System is routed over the 56K links.
table 14.1 and table 14.2 show the match and set commands that can be used with redistribution, and table 14.3 and table 14.4 show the match and set commands that can be used with policy routing. Table 14.1. Match commands that can be used with redistribution.
Table 14.2. Set commands that can be used with redistribution.
Table 14.3. Match commands that can be used with policy routing.
Table 14.4. Set commands that can be used with policy routing.
|