10.9 Summary

There are two basic types of routing protocols: interior gateway protocols (IGPs), for routers within an autonomous system, and exterior gateway protocols (EGPs), for routers to communicate with routers in other autonomous systems.

The most popular IGP is the Routing Information Protocol (RIP) with OSPF being a newer IGP that is gaining widespread use. A new and popular EGP is the Border Gateway Protocol (BGP). In this chapter we looked at RIP and the types of messages that it exchanges. RIP Version 2 is a recent enhancement that supports subnetting and other minor improvements. We also described OSPF, BGP, and classless interdomain routing (CIDR), a newer technique being deployed to reduce the size of the Internet routing tables.

There are a two other OSI routing protocols that you may encounter. Interdomain Routing Protocol (IDRP) started out as a version of BGP modified for use with OSI addresses instead of IP. Intermediate System to Intermediate System Protocol (IS “IS) is the OSI standard IGP. It is used for routing CLNP (Connectionless Network Protocol), an OSI protocol similar to IP. IS-IS and OSPF are similar.

Dynamic routing is still a fertile area of internetworking research. The choice of which routing protocol to use, and which routing daemon to run, is complex. [Perlman 1992] provides many of the details.

Exercises

10.1	In Figure 10.9 which of the routes came to `gateway` from the router `kpno?`
10.2	Assume a router has 30 routes to advertise using RIP, requiring one datagram with 25 routes and another with the remaining 5. What happens if once an hour the first datagram with 25 routes is lost?
10.3	The OSPF packet format has a checksum field, but the RIP packet does not. Why?
10.4	What effect does load balancing, as done by OSPF, have on a transport layer?
10.5	Read RFC 1058 for additional details on the implementation of RIP. In Figure 10.8 each router advertises only the routes that it provides, and none of the other routes that it learned about through the other router's broadcasts on the 140.252.1 network. What is this technique called?
10.6	In Section 3.4 we said there are more than 100 hosts on the 140.252.1 subnet in addition to the eight routers we show in Figure 10.7. What do these 100 hosts do with the eight broadcasts that arrive every 30 seconds (Figure 10.8)?

Chapter 11. UDP: User Datagram Protocol

11.1 Introduction

11.2 UDP Header

11.3 UDP Checksum

11.4 A Simple Example

11.5 IP Fragmentation

11.6 ICMP Unreachable Error (Fragmentation Required)

11.7 Determining the Path MTU Using Traceroute

11.8 Path MTU Discovery with UDP

11.9 Interaction Between UDP and ARP

11.10 Maximum UDP Datagram Size

11.11 ICMP Source Quench Error

11.12 UDP Server Design

11.13 Summary

Exercises

11.1 Introduction

UDP is a simple, datagram-oriented, transport layer protocol: each output operation by a process produces exactly one UDP datagram, which causes one IP datagram to be sent. This is different from a stream-oriented protocol such as TCP where the amount of data written by an application may have little relationship to what actually gets sent in a single IP datagram.

Figure 11.1 shows the encapsulation of a UDP datagram as an IP datagram.

Figure 11.1. UDP encapsulation.

RFC 768 [Postel 1980] is the official specification of UDP.

UDP provides no reliability: it sends the datagrams that the application writes to the IP layer, but there is no guarantee that they ever reach their destination. Given this lack of reliability, we are tempted to think we should avoid UDP and always use a reliable protocol such as TCP. After we describe TCP in Chapter 17 we'll return to this topic and see what types of applications can utilize UDP.

The application needs to worry about the size of the resulting IP datagram. If it exceeds the network's MTU (Section 2.8), the IP datagram is fragmented . This applies to each network that the datagram traverses from the source to the destination, not just the first network connected to the sending host. (We defined this as the path MTU in Section 2.9.) We examine IP fragmentation in Section 11.5.