Chapter 12: Internetworking and the Internet

1
Chapter 12Internetworking and the Internet

• Principles of Computer Networks and
  Communications
• M. Barry Dumas and Morris Schwartz

2
Objectives

• Define and explain internetworks and intranets
• Describe the Internets topology and explain why
  its structure might be described as
  pseudo-hierarchical
• Discuss the beginnings of the World Wide Web, its
  evolution and its relation to the Internet
• Describe Internet networking with the
  client/server model
• Explain the composition of URLs and examine
  addressing issues
• Discuss issues associated with IPv4 addressing
  and the move from IPv4 to IPv6

3
Overview

• Internetwork a group of autonomous networks
• Company internets and intranets typically
  revolve around LANs
• When varying locations are involved, use WANs
• Creating an InterNetwork Requires paying
  attention to
• Cost
• Compatibility
• Security
• Reliability
• The IPv4 system will soon be out of addresses
• A move to IPv6 system is necessary

For companies that form partnership
When these networks use TCP/IP protocols, theyre
called extranets
4
Overview

• Company intranet
• Company-owned, in-house network
• Uses TCP/IP protocols
• Designed only to be reached by authorized
  employees
• Company extranet
• Company-owned, special outsider access to
  in-house network
• Uses TCP/IP protocols
• Connects between the owner company and networks
  of participating organizations (e.g.,
  suppliers, outsourcers, etc.)

5
History of the Internet Revisited

• Usually traced back to its precursor, the ARPANET
  project
• Main concerninterconnecting independent
  (mainframe) computers
• Later concernthe development of a robust
  internetwork
• That could keep military communications flowing
• That could deal with complicated communications
  with incompatible networks
• Can be linked to the Advanced Research Projects
  Agency (ARPA)
• The U.S. response to the 1957 USSR launch of the
  Sputnik

6
Internet Topology and Access
Compromises of millions of interconnected hosts,
l LANS and WANS

• Service providers
• Organizations whose nodes and links supply all of
  the interconnections
• Order of main hierarchy
• International Internet service providers (IISPs)
  and national service providers (NSPs) at the top
• Most NSPs are also IISPs
• Regional service providers (RSPs)
• Local Internet service providers (ISPs) at the
  bottom

Many providers connect directly to each other,
whether at the same or different levels
Local providers offer dial-up access, bringing
the telephone system into the picture
7
Internet Topology and Access

• National service providers (NSPs)
• Form the Internet backbone that extends worldwide
• Are private companies that own and maintain the
  backbone networks
• Basic global interconnections are provided by
  NSPs linked to each other through network access
  points (NAPs)
• NAPs are complex switching stations
• NAPs are privately owned, usually by companies
  other than NSPs
• Some NSPs bypass NAPs to link directly to each
  other using peering points in their switching
  offices

Peering points are like the point of presence
POPs which is the location of a node on a network
that users can connect to. i.e telephone
companies end offices
8
Internet Topology and Access

• Regional service providers (RSPs)
• Through routers
• Connect hierarchically to NSPs
• Connect directly to other RSPs
• Local Internet service providers (ISPs)
• Can link to NSPs, RSPs, and ISPs
• The higher up on the hierarchy, the faster the
  links and the greater their capacity
• ISPs can support many connection types
• Dial-up, cable modem, DSL, ATM, frame relay,
  Ethernet
• Not all ISPs can support all types

Most individuals and businesses use ISPs to
connect
9
Basic Topology of the Internet
Some RSPs connect directly to each other by
routers
NSPs are linked to each other by NAPs
Some NSPs connect directly to each other by
peering points
Fig. 12.1
10
Internet2 and Abilene (a complete separate entity)
Will be covered in ch 18

• Internet2 1995
• Nonprofit development project
• Academic, industry, government partnership
• Led by more than 200 universities (alliance)
• Purposeto create advanced technologies and
  applications that can be adopted by the Internet
• Will eventually lead to the Internet of the
  future
• Formation and constituency go back to its
  predecessors
• Abilene
• High-speed wide-bandwidth optical backbone
  network
• Designed to support Internet2

Abilene participantsIndiana UniversityJuniper
NetworksNortel NetworksQwest Communications
in partnership with Internet2
11
The World Wide Web aka the Web

• An interface that allows us to access the
  Internet
• the Web to the Internet is the same as the
  database application to a database
• Tim Berners-Lee in 1990
• Wrote the first World Wide Web server httpd
• Created WorldWideWeb
• the first client
• a hypertext browser/editor
• Web browser software
• Simplified the information-finding process on the
  Internet providing easy-to-use Web interfaces
• Websites
• Collections of files (pages) organized by links
• Via a structure called hypertext (that contains
  hyperlinks)

Web interfacesMicrosoft Internet
ExplorerNetscape NavigatorMozilla Firefox
12
The Client/Server Model

• Client Software requests services, Servers
  Software provide services
• Name refers to the association between network
  entities
• Client software requests services
• Server software provides services
• A software model, not a hardware model
• Because it is software based, the client/server
  model provides a flexible and scalable
  architecture
• This explains its popularity
• Different from master/slave relationship!
• Server software in server/client model does not
  control the network as in the case of the master
  slave
• Servers and clients operate independently

Servers and clients only join for the
requestresponse relationship
13
The Client/Server Model

• Client/Serverhow different types of software
  running on network devices interact
• Examples
• When you go to a website, your browser software
  (client) requests Web pages from the sites Web
  server software (server)
• You can download a file from an Internet server
  by using an FTP (file transfer protocol) client
  that requests the file from a server running FTP
  software (part of the TCP/IP protocol suite)
• An application can be both a client and a server
• One time requesting services and another time
  providing them

This is common in peer-to-peer networks
14
The Challenge of Internetwork Addressing

• Standardized protocols and procedures are key
  factors in Internet success
• To send a message, the system must
• Resolve the location of the recipient machine
• Distinguish it from all the devices on the
  Internet
• Computers on a shared medium LAN (not an
  internetwork) have unique flat physical addresses
• Makes recipients easy to identify, but
• Insufficient and impractical for internetworking!
• Addresses do not contain any location information
• System would have to search every network in the
  internetwork for the recipient machine

What is the solution for this problem? See next
slide
15
The Challenges of Internetwork Addressing
solution

• Hierarchical scheme
• Different levels identify
• A particular network of the internetwork
• The physical machine address
• Two architecture models
• Open systems interconnection (OSI) model
• The medium access control (MAC) sublayer of the
  data link layer handles physical addresses
• Network layer handles logical addresses
• Transmission control protocol over Internet
  protocol (TCP/IP) model
• Follows the same pattern as OSI, but with
  possibly different labels
• OSI data link layer is the TCP/IP data link or
  link layer
• OSI network layer is the TCP/IP network or
  Internet layer

16
Hierarchical Addresses

• The postal system uses hierarchical addresses
• Zip codes, states, cities, streets, names, etc.
• Allows the post office to route mail in stages
• Hierarchical network addresses similarly comprise
  groupings/segment
• Allow the system to route messages to general
  areas, particular networks and subnetworks, and
  finally the destination machine
• Addresses are constructed and routed in network
  layer (OSI) or internetwork layer (TCP/IP)

(Reviewing from Chapter 6)
17
Hierarchical Addresses

• Physical address is different from the network
  address
• Physical addressrefers to a particular device
• The physical address doesnt change when the
  device is moved
• Network addressrefers only to the network in
  which the device resides
• The network address changes when the device is
  moved!
• Analogy
• An automobile VIN stays with the automobile
  (physical address) if you move to a different
  state
• The license plate (network address) changes to be
  state-specific

18
Addressing in the Internet
In 1983 ARPANET officially adopted TCP/IP as the
standard communications protocol.

• Replaced NCP (network control protocol)
• Major step towards todays Internet
• Explains why the Internet uses TCP/IP model
  architecture
• TCP/IP
• groups application functions into a single
  applications layer
• Communications functions are in the other layers
• OSI
• Layers above transport focus on applications
• Layers below session deal with communications

See next slide
19
Model Architectures
Fig. 12.2
20
Addressing in the Internet

• Internet protocol (IP) address
• Used to identify a device for the Internet, in
  the internet layer
• Different from a medium access control (MAC)
  address
• IP address
• Associated with a machine that may or may not be
  in a LAN
• A logical address at the internet layer
• May be changed without affecting the physical
  address
• MAC address
• A physical address at the data link layer of a
  device on a LAN

21
Addressing in the Internet

• IP address
• Can be
• Static
• Assigned and fixed on the device by a network
  administrator
• Dynamic
• Assigned to a device by a protocol process when
  the device links (logs on) to the Internet
• Dynamic IP addresses are recycledreleased when a
  device disconnects and available for assignment
  on another device
• Is used by the Internet to route packets
• To reach a device, there must be a mapping of its
  IP address to its physical address

In other words, the IP address must be associated
with the devices physical address
22
how to find the IP and the MAC address on your
computer

• For the IP address
• Run
• Cmd
• Ipconfig
• MAC address
• open the network connections
• Select your LAN connections, right click, select
  status
• In the support tap click Details
• Your MAC is the physical address

23
Addressing in the Internet

• Mapping of its IP address to its physical
  address
• There are several protocols to do this mapping
  (i.e., IP address to physical address)
• Address resolution protocol (ARP) ltlt Original
• Reverse address resolution protocol (RARP) ltlt
  companion of ARP
• Dynamic host configuration protocol (DHCP) ltlt
  new

)More about these in Chapter 13)
24
The Domain Name System

• Domain name
• The alphabet version of an IP address on the
  Internet
• Domain name system (DNS)
• Used by the internet to translate a domain name
  or e-mail address to an IP address
• Every domain name and e-mail address
• Is globally unique
• Has a one-to-one relationship with a unique IP
  address
• Resolving the domain name
• The process where DNS translates a typed domain
  name into an IP address that the Internet uses to
  route the transmission

For example, www.icann.org resolves (translates)
into dotted quad notation as 192.0.34.65
Translates into Binary 32 bits 4x8
25
The Domain Name System
The translation process is called resolving the
domain name, applies for e-mail as well

• E-mail addresses
• A computer program called a mail transfer agent
  sends e-mail from one computer or mail server to
  another
• These agents use the DNS to find out where to
  deliver the email
• Smooth operations in the DNS
• DNS is an interconnected hierarchical system of
  high-speed servers running distributed domain
  name databases
• For translation, this system simply searches its
  databases, finds the IP address for the name, and
  relays it back
• Centralized organization keeps the DNS up to date
  (new additions or deletes)

Domain name registries are responsible for
distributing domain names and IP addresses while
ensuring their uniqueness
26
The Parts of a URL

• Uniform resource locator (URL)
• Is a symbolic meaning for specifying
• a Web resource
• The Web server on which the resource resides
• The protocol that will be used to retrieve the
  resource
• URL components are separated from each other by
  forward slashes, dots, and sometimes colons
• Easiest to interpret from right to left
• The rightmost segment is called the top-level
  domain (TLD)

27
Top-Level Domains (TLDs)

• Easier to interpret if starting from right to
  left
• www.users.alvernia.edu
• Five original TLDs
• .com for commercial enterprises
• .gov for government sites
• .net for organizations providing network services
• .mil for use by the military
• .org for nonprofit organizations and those that
  do not fit other designations
• Because .com, .org, and .net characteristics have
  blurred over time, they are now referred to as
  generic TLDs (gTLDs)
• TLD concept speeds up the searching process in
  the database because each partition is
  relatively small

TLD
28
Domain and Sub-domain Names

• Domain name
• www.users.alvernia.edu
• Also called second-level domain
• To the left of the TLD, separated by a dot
• Specifies a particular network, an autonomous
  system (AS) within the Internet
• Sub-domain name
• www.users.alvernia.edu
• Narrows the location of the resource server

29
URL Server

• Server (host) name
• www.users.alvernia.edu
• Is located to the left of the sub-domain name
• Holds the requested resource

It is common practice to give the name www to the
server that hosts Web documents
However, it is not required!
30
Domain Name and URL Components
Combined domain name.cuny.edu specifies a
particular network within the Internet
www is a server at Baruch College
Fig 12.3
If you see a URL that ends after the TLD or after
a subdirectory name, the extension/index.htm or
/index.html is assumed
31
Specifying the File on the Server

• Domain names
• Specify location of the server
• Do not explicitly specify the file (Web page) on
  the server
• Beyond domain names
• We need the path to the file on the server
• Path must include directories and the file name
• Path information is appended to the right of the
  TLD by a slash (/)
• Example
• www.users.alvernia.edu/students/finalgrades/index
  .htm
• /students is the directory where Web files for
  students are stored
• /finalgrades is the subdirectory where files
  specific to final grades are stored
• /index.htm specifies one particular file

32
Specifying the File on the Server

• .htm and .html
• Indicate that the file is written in hypertext
  markup language (HTML)
• Are default file names that are automatically
  searched for if no file name is given

Any URL with nothing after the TLD or a
subdirectory name assumes the extension
/index.htm or /index.html
33
Specifying the File on the Server

• The URL must inform the server of the protocol
  the client will use in the interchange process
• Specifying the protocol in the URL
• Leftmost segment of the URL defines actions taken
  in response to particular requests
• http// is one of the most common Web protocols
• Stands for hypertext transfer protocol
• In a browser, sends a command to the sites Web
  server to download the page
• Part of the application layer of the TCP/IP suite
• A stateless protocol
• Each command is performed independently
• Makes it difficult to create sites that interact
  with users

34
The Http Protocol and Cookies

• Software like Java is used to overcome
  stateless protocol difficulties
• Used to write very small text files (cookies) to
  the clients hard drive
• Cookies contain state information
• Allow a server application to understand the http
  requests that make up a continuous exchange
• http does not prevent unauthorized accessing see
  next slide

35
Other Identifiers (common protocols)

• https//
• For sites that require secure transmissions, an s
  is added, indicating encryption
• Unreachable without appropriate passwords
• ftp (file transfer protocol)
• Commonly employed protocol
• Used for uploading and downloading files to and
  from ftp servers
• ftp is typically in the server name, but not
  required
• Country identifier
• The country identification is part of the TLD,
  though separated from it by a dot
• For example, BBC News has a United Kingdom
  identifier
• news.bbc.co.uk
• When with the TLD, it is called a country code
  top-level domain (ccTLD)

There are more than 240 ccTLDs!
36
IPv4

• IP addressing began with ARPANET
• 1981 IPv4 became the standard we use today
• Hierarchical scheme
• Classes of addressesThree logical
  arrangements/splits of the bits reserved for
  addresses
• For few organizations needing many host addresses
• Few bits for network addresses, many for hosts
• For many companies with many more hosts
• Many bits for network addresses, but also many
  for hosts
• For the great many organizations with very few
  hosts
• Many bits for network addresses, few for hosts

37
IPv4 Classful Addressing

• Classfulmost widely used type of IPv4
• Consists of 32 bits arranged in the dotted quad
  format
• Four 8-bit sections
• Makes up three unicast classes
• Unicastfrom one source to one destination
• Two-part addresses that split the 32-bits into
  network/host
• Class A 8 / 24
• Class B 16 / 16
• Class C 24 / 8
• Class identifier bits (prefixes) are included in
  the network address part of the split

192
.0
.34
.65
38
Classful Addressing Prefixes

• Prefixes
• Identify class
• Are not part of the IP address
• Class A is 0
• Class B is 10
• Class C is 110
• D (not classful) is 1110 used for multucasting
• E (not classful) is 1111 for Expermental

Starting bit
39
IPv4 Classful Addressing
From A to E Networks increase and hosts decrease
These classes account for 87.5 of potentially
available addresses
Class Prefix (1st 8-bit section) Number of Networks Number of Hosts
A 0 _ _ _ _ _ _ _ 27 2 126 224 2 16,777,214
B 10 _ _ _ _ _ _ 214 2 16,382 216 2 65,534
C 110 _ _ _ _ _ 221 2 2,097,150 28 2 254
Table 12.1
40
IPv4 Non-Classful D and E

• Two other categories of bits reserved for
  addresses
• D and E are not segmented into networks and hosts
• Both allow for 228 268,435,456 addresses
• D
• Multicasting
• From a source to multiple destinations
• E
• Reserved for experimenting

41
Class A address Network Address

• 32 bits
• First 8 left most for network address
• The other 24 bit for the host
• First left most bit used as a class identifier
• No address can be all 1s or all 0s
• For 8 bits 2n 27 gives 128 address
• Without the address of all 1s or 0s we get 126
  network addresses

42
Class A address host Address

• 24 bits
• First 24 right most for host address
• No address can be all 1s or all 0s
• For 24 bits 2n 224 gives 16,777,216 address
  
• Without the address of all 1s or 0s we get
  17,777,214 host addresses
• Same calculations for class B and class C

43
Classful Addressing

• An organization that applies for an IPv4 address
• Receives a network address with a block of host
  addresses
• The size of this block is determined by class
• If the organization can handle more addresses
  than it actually uses, the other addresses
  associated with the companys block go unused
• Significant limitation to classful addressing
• It wastes a lot of addresses

Soon they will run out of addresses!
To forestall this, classless addressing was
implemented
44
Classful Addresses, Networks, Subnets, and Masks

• Network ID
• A company receives a network ID when a classful
  network address is assigned
• Network ID host address all 0s network
  address
• Used by outside routers to direct IP packets
  addressed to the company
• Not assignable to any company host
• No host ID can be all 0s
• Logical IP networks
• A company subdivides the classful network address
  to organize its own hosts

45
Subnets and Masks

• Subnets
• Logical networks with their own subnet addresses
• Created by assigning hosts to groups with their
  own subnet addresses
• Organized many waysby building, floor,
  department, LAN
• Major advantages
• Better control on subdividing and managing the
  network
• Masks
• Bit patterns applied to entire addresses to
  isolate their components
• Used to separate network, subnet, and host
  addresses
• Have the same number of bits (arranged in dotted
  quad segments) as the IP address, but only use 1s
  and 0s

A single IP address can connect a whole subnet
to the Internet
46
Bitwise Multiplication and Masks

• Bitwise multiplication of the address by the mask
• Equivalent to applying the and operator
• Captures address parts where mask bits are 1 and
  ignores where they are 0
• Internet routers easily identify the IP address
  class by finding bit patterns this way

1st the class is identified by checking the left
most bits, then the prober mask is applied
Class B mask
47
Bitwise Multiplication and Masks

• When the class is identified, a network default
  mask is applied
• Three default masks
• Class A mask 255.0.0.0
• Class B mask 255.255.0.0
• Class C mask 255.255.255.0

48
Bitwise Multiplication and Masks

• In operation
• After one of the three default masks is applied,
  the network address is revealed
• The network address is assigned to the edge
  router of the organization
• When a packet reaches any router, the appropriate
  mask is applied
• If the network address it finds is not for that
  organization, the packet is passed to the next
  hop router
• If the network and router addresses match, a
  subnet mask is applied

49
Addressing in the Internet

• Subnet address
• Comprises the network address subnet mask bits
• The remaining host address bits are all 0s
• The total number of bits in the combined network
  and subnet addresses is indicated by a /n
  notation at the end of the address

130.57.110.9/19
19 bits
16 bits
3 bits
50
Classless Addresses

• A solution to the IP address shortage?
• Classless addressing
• All of IPv4s address space of 32 bits would be
  available without restriction
• Twice as many addresses could be created
• But addressing hierarchy and restrictions needed
• Otherwise, routers would be overwhelmed and
  complicated

51
Classless Addresses

• Classless inter-domain routing (CIDR)
• The compromise between classful and classless
• Allows any number of leftmost bits to be assigned
  as a network address
• Addresses assigned based on the number of hosts a
  network can support no class designation
• CIDR is not limited to network addresses of
  8,16,or 24 bits
• CIDR is NOT perfect
• Still wastes addresses, just not as many as
  classful addressing

52
CIDR, Subnetting, and Supernetting

• Supernetting
• CIDRs hierarchical scheme that parallels
  subnetting
• One key differenceit is applied to routing
  outside of the organization (hence the name)
• Is a method of route aggression
• A single high-level routing table entry
  represents many lower-level routes
• Internet backbone routers need fewer entries
• More efficient, eases table size requirements

53
IPv6

• Even with CIDR, supernetting and subnetting,
  still address shortage
• IPV6 replaced IPV4
• Uses a 128-bit address sequence instead of 32
• Provides IP header extensions
• Adds quality of service (QoS) labeling to IP
  packets
• Uses coloned octal, not dotted quad
• Accommodates CIDR by adding a (an) /n to the end
  of the address

54
IPv6

• Uses a 128-bit address sequence instead of 32
• increases the number of available IP addresses
• allows for additional hierarchy levels that
  improve routing efficiency
• Provides IP header extensions
• Adds quality of service (QoS) labeling to IP
  packets
• Uses coloned octal, not dotted quad
• Accommodates CIDR by adding a /n to the end of
  the address

55
IPv6

• Uses a 128-bit address sequence instead of 32
• Provides IP header extensions
• Improve privacy, authentication, and integrity
• Adds quality of service (QoS) labeling to IP
  packets
• Uses coloned octal, not dotted quad
• Accommodates CIDR by adding a /n to the end of
  the address

56
IPv6

• Uses a 128-bit address sequence instead of 32
• Provides IP header extensions
• Adds quality of service (QoS) labeling to IP
  packets
• Specifies the level of service requests
• Priority, real-time, normal handling
• Uses coloned octal, not dotted quad
• Accommodates CIDR by adding a /n to the end of
  the address

57
IPv6

• Uses coloned octal, not dotted quad
• Eight segments separated by colons
• A1B9CC5F000D0037FF0E394500002A4D
• Two bytes per segment
• Typically written in hexadecimal notation
• Still 32 characters, one hexadecimal digit 2
  bytes
• Leading 0s in each section are eliminated for
  simplification
• A1B9CC5F000D0037FF0E394500002A4D
• A1B9CC5FD37FF0E394502A4D
• BUT, only one string of 0s can be removed in a
  given address

58
IPv6

• Uses a 128-bit address sequence instead of 32
• Provides IP header extensions
• Adds quality of service (QoS) labeling to IP
  packets
• Uses coloned octal, not dotted quad
• Accommodates CIDR by adding a /n to the end of
  the address
• n is the number of bits in the CIDR prefix

59
IPv4 Packet Headers
Fig. 12.4IPv4
60
IPv6 Packet Headers
Fig. 12.4IPv6
61
Moving from IPV4 to IPV6

• A huge difference between both methods
• Has to be done gradually
• Three methods to allow gradual transition and to
  permit functioning in mixed environment

62
Methods for Moving from IPv4 to IPv6

• Dual stack
• What?
• Stackthe IP protocols used by the network nodes
  (routers, hosts)
• Dual stacknodes that contain the stacks for both
  IP versions
• How?
• The sender queries the DNS for an address
• If the address is IPv4, the packet is sent as
  IPv4
• If the address is IPv6, the packet is sent as
  IPv6
• Once the change to IP6 is complete IP4 can be
  deleted
• Pro
• Network nodes accommodate both IPv4 and IPv6
• Con
• Each of the dual stack nodes must have an IPv4
  address
• Address scarcity is not alleviated
• Processing through two stacks adds to switching
  time

63
Transitioning from IPv4 to IPv6
Both IPv4 and IPv6 addresses are maintained
The sender uses whatever packet format (i.e.,
IPv4 or IPv6) is returned from the DNS server
for the destination node
Fig. 12.5A
64
Methods for Moving from IPv4 to IPv6

• Tunneling
• Why?
• A packet from an IPv6 node or region of nodes (a
  cloud) may have to travel across an IPv4 cloud to
  reach another IPv6 node
• How?
• An IPv4 tunnel is created for it to travel
  through
• First it needs an IPv4 address from the IPv6 edge
  router at the IPv4/IPv6 border
• The IPv6 router will encapsulate it into an IPv4
  packet
• At the other border, the IPv4 edge router will
  then decapsulate this packet
• Pro
• Avoids having to assign IPv4 addresses to
  IPv6-only nodes within a capsule
• Con
• Additional processing at the borders

65
Transitioning from IPv4 to IPv6
An IPv4 header encapsulates IPv6 packets while
transiting through IPv4 regions
Fig. 12.5B
66
Methods for Moving from IPv4 to IPv6

• Translation
• Why?
• An IPv4-only host cannot understand packets from
  a IPv6-only host
• Tunneling will not help resolve this problem
• The packet is still IPv6 after the encapsulating
  header is removed
• How?
• At the least, the edge router must translate the
  IPv6 header into an IPv4 header
• Pro
• IPv4 hosts and IPv6 hosts can communicate
• Con
• Translation can be complicated!
• The end node processes can involve the IP
  protocols themselves