IP Routing and Forwarding

IP has emerged as the dominant Layer 3 protocol for connectionless networking. IP is part of a suite of protocols referred to as the Transmission Control Protocol/Internet Protocol (TCP/IP) protocol suite (or simply as the Internet Protocol suite). The IP suite embraces many protocols, some of which are listed in RFC 1800, "Internet Official Protocol Standards." This list of IP- related protocols continues to grow as new applications emerge. Some key IP protocols include the following:

  • Internet Protocol (IP)

  • Transmission Control Protocol (TCP)

  • User Datagram Protocol (UDP)

  • Internet Message Control Protocol (ICMP)

  • Address Resolution Protocol (ARP)

TCP and UDP provide transport for applications and run over the IP layer. ICMP is a control protocol that works alongside IP at the network layer. ARP provides address resolution between the network layer and the underlying data link layer. Numerous applications use the transport services of TCP and UDP. Some common examples include the following:

  • Telnet ” A virtual terminal application that uses TCP for transport

  • File Transport Protocol (FTP) ” A file transfer application that uses TCP for transport

  • Trivial File Transfer Protocol (TFTP) ” A file transfer application that uses UDP for transport

  • Domain Name Service (DNS) ” A name-to-address translation application that uses both TCP and UDP transport

IP is largely responsible for the continuing success of the TCP/IP suite. The popularity of IP is mainly centered on its simplicity and high efficiency for data transfer. As a connectionless protocol, IP forwards data in self-contained routable units known as datagrams or packets. Each packet contains information, such as source and destination addresses, which is used by routers when making forwarding and policy decisions.

In connectionless networking, there is no need for prior setup of an end-to-end path between the source and destination before data transmission is initiated. A file can be transmitted from one end of the network to another by breaking it down into packets, each of which is forwarded independently along the best path by routers located between the source and destination. IP forwarding is primarily based on the destination address, even though the source address and other parameters in the IP header can be used for policy-based forwarding.

IP forwarding is, therefore, commonly referred to as destination-based. Routing and forwarding essentially mean the same thing with regard to IP, even though they've taken different shades in meaning along with the evolution of routers. A router is a network device that essentially consists of a collection of network interfaces linked together by a high-speed bus or a complex interconnection system, such as a crossbar-switch or shared memory fabric. A router has two functional planes: data and control. Frequently, both functions are performed by an intelligent subsystem known as the route processor. Most modern high-speed routers are designed with a clear separation between the control and data planes (see Figure 1-1). The control plane functions are centered on building the necessary intelligence about the state of the network and a router's interfaces. IP routing applications or protocols provide the framework for gathering this intelligence. The data plane handles actual packet processing and forwarding by relying on the intelligence of the control plane.

Figure 1-1. Distributed router architecture.

graphics/01fig01.gif

IP routing is the broader process of collecting routing information about the network, a function that is performed in the control plane. IP routing protocols process this information to determine the best paths to known destinations in the network. The known best paths are stored in the routing table or the routing information base ( RIB ). The routing table is then used for forwarding packets, moving them out of the router onto the best paths to the next hop and toward their intended destinations. The best path is frequently the path with the lowest value of metric or cost to the destination.

IP forwarding involves processing information in the header of an IP packet to determine how to advance it toward the target destination. This includes activities such as looking up the destination address in a forwarding database for the exit interface, reducing the IP time-to-live (TTL) value, calculating the IP checksum value, queuing the packet at the exit interface, and eventually getting the packet out of the router onto the link to the next-hop router. Similar forwarding functions are performed independently on the router at the next hop and at every router in the path, each time getting the packet closer to its destination until it finally arrives there.

Figure 1-1 illustrates the architecture of a router with distributed forwarding capabilities. In this architecture, each interface processor (or line card) features an independent forwarding engine, which is responsible for IP forwarding. The interface processors directly switch packets between each other. IP forwarding engines are optimized for faster packet processing and switching from the source interface to the destination interface on a router.

The IP routing functionality is performed on the route processor, which has a routing engine for calculating routes. The route processor runs a routing protocol that allows it to interact with other routers, gather and process routing information, and build the routing table. The route processor is optimized for gathering routing information, which it eventually shares with the interface processors. The Cisco 12000 Series routers have a fully distributed architecture.

Figure 1-2 shows an alternative router architecture with a dedicated packet processor featuring a high-speed forwarding engine designed to provide centralized packet switching for the whole system. Also shown is a separate route processor. As in the distributed architecture, the data and control planes are separated. Interface processors in the centralized architecture do not have forwarding intelligence to exchange packets directly but, instead, direct all packets to the packet processor where actual forwarding is done. This type of router architecture is described as centralized.

Figure 1-2. Centralized router architecture.

graphics/01fig02.gif

Fully distributed and centralized router designs are the extremes of router architecture options, and there are various hybrid options in between them. One option is to integrate the packet and route processors shown in Figure 1-2 into a single hardware element, commonly referred to as a route switch processor. Other options mix centralized and distributed forwarding capabilities in the same router as in the architecture of the Cisco 7500 Series routers. These routers feature hardware modules called versatile interface processors ( VIP ) and route switch processors ( RSP ). The VIP provides distributed forwarding, whereas the RSP combines routing and centralized packet-processing capabilities.

The routing information collected by a router and shared with other routers in the network consists of IP subnets or address prefixes that are associated with various links in the network. The hosts in a network where most applications reside are typically connected to local-area network (LAN) media. The IP address of a host is based on the subnet assigned to its LAN.

The following section briefly discusses IP addressing in general and provides a refresher for IP subnetting and related subjects, such as variable-length subnet masking ( VLSM ) and classless interdomain routing ( CIDR ). Later discussions focus on various categories of IP routing protocols and packet-switching mechanisms used for IP forwarding on Cisco routers.



IS-IS Network Design Solutions
IS-IS Network Design Solutions (Networking Technology)
ISBN: 1578702208
EAN: 2147483647
Year: 2005
Pages: 144
Authors: Abe Martey

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