Summary

This broad chapter on IP routing concepts provides readers with an overview of IP routing protocols and an understanding of the underlying mechanisms involved in routing packets generally in any IP network and specifically on Cisco routers.

Addressing is a key aspect of routing, and IP addressing bears similar importance in IP routing. Subnetting and classless notions in IP addressing are intended to provide efficient allocation schemes, which effectively result in efficient use of assigned addresses from the public IP address space, as well as conservation of network resources in maintaining IP routing information on routers. Classless addressing provides the framework for prefix-length-based IP addressing schemes, such as CIDR and VLSM. Various types of IP routing protocols exist, differing in protocol design as well as functional capabilities. In general, routing protocols can be classified into various categories, as follows : unicast versus multicast, classful versus classless, intradomain versus interdomain, and distance-vector versus link-state protocols. These classifications provide insights into the capabilities of well-known routing protocols, emphasizing strengths and weaknesses that should be considered when selecting the appropriate routing protocol for a network design project.

The subtle distinction between IP routing and IP forwarding is important for understanding the architecture of modern routers. IP routing extends the notion of IP forwarding to include the process of gathering routing intelligence in a network. The concept of next -hop routing, using the destination address in packets, is the basic forwarding paradigm of IP, even though modern routers can use policy-based mechanisms as well, based on the source address for forwarding. The following three basic IP forwarding methods are supported in Cisco IOS Software: process switching, fast switching, and CEF. Process switching is CPU-based forwarding, and it is much slower than the other switching methods . Fast switching uses a demand-based forwarding cache to speed up the switching process, whereas CEF optimizes further by precomputing the information used for switching packets over the whole of the routing table.

IP networking is still evolving with IPv6 being a fairly recent introduction that is not yet widely accepted. IPv6 provides enhancements to the IPv4 architecture by addressing some of the limitations of the latter, such as the size of the IPv4 address space and problems with renumbering networks. IPv6 addresses are 128 bits long compared to the 32 bits of IPv4 and should provide a large enough address space for further growth of the global Internet.