1 Managing IP Addresses and Broadcasts Chapter 2.

1

Managing IP Addresses and Broadcasts

Chapter 2

22

Making Networks Scalable

A scalable network grows continually, yet smoothly and stably

Avoid problems with growing networks by providing redundancy and designing networks for easy manageability

Choice of routing protocol greatly influences scalability of network

33

The Growth of the Internet

Initially, Internet was small and limited to researchers

In 1990s, Internet grew immensely as governments, universities, corporations, and the general public began to use it

Organizations and Internet now experiencing problems managing IP addresses

44

IP Address Exhaustion

32-bit IP addresses provide, in theory, over four billion addressesMany allocated addresses are wasted

Fear that the Internet may run out of usable IP addresses

55

Wasting Addresses

It is acceptable to use subnet zero and the all-ones subnet with VLSM.(In the past, use of the first and last subnets was discouraged).

Consider the following example:

In this network a Class C address with a 255.255.255.0 mask has been used for each subnet

192.168.1.0/24

192.168.2.0/24

192.168.3.0/24

The WAN link has enough IP addresses for 254 separate hosts, but will use only two. Each LAN has enough IP addresses for 254 separate hosts. Broadcasts would be a major issue if this address space were not further subnetted.

Consider this alternative addressing scheme:

192.168.0.0/25 192.168.0.128/26

192.168.0.192/30

This network allows 126 different host addresses

This network allows 62 different host addressesThis network allows just

2 host addresses

66

Routing Table Growth

Internet routing table increased from about 5000 routers in 1990 to more than 100,000 in 2001

Large routing tables require more CPU time and more memoryResult in slowed down table lookupsMake troubleshooting more difficult

77

Managing IP Addresses

Administrators use many strategies to manage IP addresses

Hierarchical addressing Hierarchical routing Route summarization Variable-length subnet masks Classful and classless routing

88

Hierarchical Addressing

Layered, orderly addressing Similar to public telephone network

Local office recognizes local exchangeLocal central office forwards long distance

calls to central office in other area codesCalls then treated as local call by

central office in other area codes

99

Hierarchical Routing

• Router forwards packet to core layer router based on first octet IP address

• Core layer router forwards packet to distribution layer router based on first two octets

• Distribution layer router forwards packet to access layer router based on first three octets

• Access layer router forwards packet to final destination

1010

Route Summarization

Also called address aggregation Combines multiple routes that share

leftmost bits into one summary routeSimilar to telephone area code

Reduces number of routes to a specific customer

1111

Route Summarization

INSERT FIGURE 2-2

1212

Route Summarization

• If router has both summary route and ordinary route, it selects the one with the longest matchLooks at length of prefix or number of bits in

subnet mask to determine path• Route summarization does not make

address allocation more efficient, especially point-to-point links

1313

Example of Routing Table with Multiple Routes to a Destination

1414

Without Route Summarization

1515

With Route Summarization

1616

Variable-Length Subnet Masks

VLSMs, defined in RFC 1812, let you subdivide Class C

Subnet mask helps router break IP address into network and host portions Router uses network part of IP address to

forward packet to correct networkLocal router uses host part of IP address to

deliver packet to destination

1717

Example of Calculating the Network Number

INSERT FIGURE 2-4

1818

The Logical AND Operation

• Router matches bits in IP address and subnet mask

• Compares bits and performs logical AND operation If both bits are ones, the result is a one If either bit is a zero, the result is a zero

• Logical AND operation provides network number

1919

Example of Logical AND Operation

INSERT TABLE 2-1

2020

Calculating Subnets

Number of subnets depends on number of bits borrowed from network portion of IP address

Calculate number of new subnets by 2n, where n is the number of borrowed bitsSubtract two to find number of usable host bits First and last addresses reserved for network

address and broadcast address

2121

Classful and Classless Netmasks

If netmask follows traditional class boundaries, it is called classful routing

If netmask does not follow traditional class boundaries, it is called classless routingCan supernet or use a smaller netmask than

traditional class boundaries

2222

Example: Calculating Subnets with VLSM

28 hosts

60 hosts 12 hosts 12 hosts

A class C address of 192.168.10.0/24 has been allocated.

Requirement levels, listed from the largest to the smallest:

Host address range

Singapore to KL

Sydney to KL

Perth to KL

12Singapore

12Sydney

28KL LAN

60Perth LAN

4th Octet network/host bitsHostsNetwork

192.168.10.0/26

192.168.10.64/27

192.168.10.96/28 192.168.10.112 /28

192.168.10.128 /30

192.168.10.132 /30

192.168.10.136 /30

2

2

2

.NNHHHHHH /26 ( 62 hosts) 192.168.10.1 - 192.168.10.62

.NNNHHHHH /27 ( 30 hosts) 192.168.10.65 - 192.168.10.94

.NNNNHHHH /28 ( 14 hosts) 192.168.10.97 - 192.168.10.110

.NNNNHHHH /28 ( 14 hosts) 192.168.10.113 - 192.168.10.126

.NNNNNNHH /30 (2 hosts) 192.168.10.129 - 192.168.10.130

.NNNNNNHH /30 (2 hosts) 192.168.10.133 - 192.168.10.134

.NNNNNNHH /30 (2 hosts) 192.168.10.137 - 192.168.10.138

2323

Calculating VLSM Subnet Masks

According to RFC 1812, all bits in subnet mask must be contiguousCisco IOS displays error message if subnet has

discontiguous bits Be sure routing protocol supports VLSMs

OSPF and EIGP support VLSMsRIP version 1 and IGRP do not support VLSMs

2424

Cisco IOS Error Message for Subnet with Discontiguous Bits

2525

Summarizing Routes Using VLSMs• VLSMs allocate IP addresses more

efficiently• VLSMs provide more flexibility in

summarizing routes Based entirely on higher-order bits they share

on the leftRoutes do not have to be contiguousPrefix of summary route based on bits shared

by all routes

2626

Route Summarization

2727

Network Numbers with VLSM

2828

Example: Route Aggregation with VLSM

200.199.62.0 /25200.199.62.128/25200.199.63.0 /25200.199.63.128/25

200.199.48.0/24200.199.49.0/24200.199.50.0/24200.199.51.0/24

200.199.32.0/22200.199.36.0/22200.199.40.0/22200.199.44.0/22

Advertise one supernet route: _______________ to RTZ

Advertise one supernet route: _______________ to RTZ

Advertise one supernet route: _______________ to RTZ

Advertise one supernet route: _______________ to ISP

200.199.62.0/23

200.199.48.0/22

200.199.32.0/19

200.199.32.0/20

2929

Classes of IP Addresses

Class depends on first octet of IP addressClass A addresses begin with a zero as the leftmost

bit; use 8 bits for network addressClass B addresses begin with a 10 as the first two

bits; use 16 bits for network addressClass C addresses begin with a 110 as the first

three bits; use 24 bits for network addressClass D addresses are used for multicastClass E addresses are used for research

3030

Classful Routing

Router uses classes of addressesCan subnet along class octet boundaries

Routing protocols include RIPv1 and IGRPMay use IP classless global configuration

command to forward packets to a summary route

Classful routing is inflexible, limited, and sometimes wasteful

3131

Classful Address Distinctions

3232

Classless Routing Ignores traditional class boundaries Protocols include OSPF and EIGRP

Can allocate and receive IP addresses as necessary Previously Three Regional Internet Registries (RIRs)

now Five, allocate IP classless addresses in blocks American Registry for Internet Numbers (ARIN) Réseaux IP Européens Network Coordination Centre

(RIPE NCC) Asia Pacific Network Information Center (APNIC) Regional Latin-America and Caribean Address Registry

(LACNIC)-2002 African Network Information Centre (AfriNIC)-2005

3333

Classless Inter-Domain Routing (CIDR)

RIRs assign addresses based on Classless Inter-Domain Routing (CIDR)CIDR discussed in RFCs 1518, 1519, and

2050 Each CIDR block has a prefix or IP

address and a prefix length or subnet mask

3434

Allocating IP Addresses

How IP addresses are allocated affects how well network performs

Pitfalls of route summarizationRequires more planningMore useful with classless routing protocolCan lead to poor path selectionCan create problem with discontiguous subnets

3535

Problems with Summarization and Discontiguous Subnets

Route summarization hides details of network from routers

Discontiguous subnets may result in outage or inability to deliver packets

3636

Discontiguous Subnets

3737

Outage Created by Discontiguous Subnets

3838

Allocating IP Addresses Using VLSMs

• Efficient allocation of IP addresses requires Allocating enough IP addresses to each subnet for

future growth Not allocating more than necessary for each subnet

• Plan for route summarization Do not assign IP addresses haphazardly Assign IP addresses based on topology

of network

3939

Example of IP Address Allocation Based on Topology

4040

Process of Assigning IP Addresses

After finding baseline subnet, calculate the number of subnets you can useCisco recommends allocating addresses from

the lowest to the highest for easier summarizing of routes

Put your largest networks into the lower subnets

4141

Other Addressing Strategies

Unnumbered interfaces Private address space Network address translation IP version 6

4242

Unnumbered Interfaces

• Configure IP on interface without explicitly using an IP address Use ip unnumbered command to refer to an existing

interface that routers use as source address Unnumbered interfaces often get IP address from

loopback address• Drawbacks include inability to get status by pinging,

making troubleshooting and monitoring more difficult• Some serial protocols such as X.25 and SMDS do

not support unnumbered interfaces

4343

Private Address Space

RCF 1918 sets aside three ranges of IP addresses for private networks 10.0.0.0/8192.168.0.0/16172.16.0.0 through 172.31.255.255

Do not route addresses in these blocks to the Internet

4444

Network Address Translation

• NAT involves device such as a router that translates one set of IP addresses into another setCan conserve IP addresses by translating a large

pool of private addresses into a small pool of public addresses

• Disadvantages include increased latency and difficulties with protocols or applications that put IP address in data portion of IP packet

4545

IP Version 6

IPv6, specified in RFC 2460, offers several advantages over current version (IPv4)Uses 128 bit IP addressesProvide over 3 x 1038 possible IP addresses Includes more support for quality of service

and better security Adoption of IPv6 is moving slowly

4646

Managing Broadcasts

Routers do not, by default, forward broadcasts If PC boots without knowing its IP address, it

must contact DHCP or BOOTP server If server not on same segment, PC cannot get an IP

addressCan hard code all IP addresses if PC unable to

reach server Creates administrative nightmare

4747

Using a Helper Address Solution is to allow broadcasts in specific

situations Cisco routers can direct a broadcast to a helper

address Can configure more than one helper address Must use IP directed-broadcast interface

configuration command with Cisco IOS 12.0 and later

Configure helper address to router closest to client By default, helper address command turns on eight

UDP ports as shown in Table 2-8

4848

Default UDP Ports