Computer Networking

1
Computer Networking
2
Bits and Bytes

• Putting information into a form that a computer
  can deal with
• A 01000001
• B 01000010

3
Information Encoding

• 065 01000001 A
• 066 01000010 B
• 067 01000011 C
• 068 01000100 D
• 069 01000101 E
• 070 01000110 F
• 071 01000111 G

4
Review (maybe)

• Have a bit
• 0 or 1
• Take a whole byte
• Eight bits
• R represents a letter or numeral or punctuation
  mark

5
Transmission of Information

• Bandwidth
• Bits per second
• Kilo
• Mega
• Giga

6
A Computer Network

• What is a computer network?
• A network is a collection of computers or
  computer-like devices that can communicate across
  a common transmission medium.

7
A Computer Network

• In a network, requests and data from one computer
  pass across the transmission medium (which might
  be a network cable or a phone line) to another
  computer.
• Example four node network

8
A Computer Network

• A computer interacts with the world through one
  or more applications (software) that perform
  specific tasks and manage input and output.
• If that computer is part of a network, then some
  of those applications must be capable of
  communicating with applications on other network
  computers.

9
A Computer Network

• A network protocol is a system of common rules
  that helps to define the complex process of
  transferring data. The data travels from an
  application on one computer, through the
  computers network hardware, across the
  transmission medium to the correct destination,
  and up through the destination computers network
  hardware to a receiving application.

10
Computer Network
Transmission medium
11
A Computer Network

• A network is usually described as being a local
  area network (LAN) or a wide area network (WAN)

12
Local Area Network (LAN)

• Many types of LAN technologies have existed over
  the years
• One predominant LAN technology exists today -
  Ethernet

13
Ethernet

• Contention media access method
• Allows many computers on the same network to
  share the same bandwidth (basically share a
  common medium or connection)
• Easily scalable easy to improve and incorporate
  new technology as it becomes available

14
Ethernet

• Uses Carrier Sense Multiple Access with Collision
  Detect (CSMA/CD)
• CSMA/CD is a protocol designed to allow multiple
  computers to share the network medium
  successfully
• Designed to manage collisions

15
Ethernet

• What is a collision?
• (example of four node 10Base-2 network)
• All computers share the connection
• Only one can transmit at a time
• Suppose computer C is transmitting information to
  computer D
• C takes over the wire sends electrical
  signals onto the wire

16
Ethernet

• All computers on the network detect the
  transmission
• Only D will process the transmitted data because
  C has addressed the information to D
• A collision will occur if two computers attempt
  to transmit at the same time (like a group of
  people talking at a party)

B
17
Ethernet

• CSMA/CD if a transmitting computer detects
  another computer attempting to transmit, it sends
  out a long jam signal that causes all computers
  on the network to be silent
• A back off scheme is used to figure out who
  gets to transmit first

18
Ethernet

• On a busy Ethernet network collisions can be a
  big problem
• SLOW!

19
Types of Ethernet

• Ethernet was initially developed by Digital
  Equipment Corporation, Intel, and Xerox
• The IEEE took their design and created the
  official network standard
• The IEEE called this standard 802.3
• 802.3 is the family name for all wired Ethernet
  types

20
Types of Ethernet 10Base2

• 10Mbps
• Baseband technology
• 185 meters (length) almost 200 meters
• 30 devices per segment
• Uses coaxial cable (coax), BNC and T-connectors
  to connect to a network
• Referred to as thinnet

21
Types of Ethernet 10Base5

• 10Mbps
• 500 meters (length)
• Up to 2500 meters with repeaters
• Up to 1024 devices for all segments
• Uses a large (thick) coaxial cable
• Referred to as thicknet

22
Types of Ethernet 10BaseT

• 10Mbps
• Uses Category 3 UTP wiring (phone wire)
• Each device connects to a hub or switch
• Only one device per segment (or wire)
• Uses RJ-45 connectors
• Supports a star topology

23
Types of Ethernet 100BaseT(X)

• 100Mbps
• Uses Category 5,6, or 7 UTP wiring
• Up to 100 meters (length)
• Only one device per segment (or wire)
• Uses RJ-45 connectors
• Supports a star topology

24
Types of Ethernet 100BaseFX

• 100Mbps
• Uses fiber optic cabling
• Up to 412 meters (length)
• Used for point-to-point connections
• Uses ST or SC connectors

25
Types of Ethernet 1000BaseT

• 1000Mbps
• Up to 100 meters (length)
• Category 5, 6, or 7 UTP wiring
• Only one device per segment (or wire)
• Uses RJ-45 connectors
• Supports a star topology

26
Types of Ethernet 1000BaseSX

• 1000Mbps
• Uses fiber optic cabling
• Up to 550 meters (length) depending upon the size
  of the cable
• Uses a 850 nanometer laser
• Uses ST or SC connectors

27
Types of Ethernet 1000BaseLX

• 1000Mbps
• Uses fiber optic cabling (multi-mode or
  single-mode)
• Up to 10 kilometers depending on type of cable
  used
• Uses a 1300 nanometer laser

28
Ethernet Addressing

• Media Access Control (MAC) address is stored on
  every Ethernet network interface card
• 48 bits long (6 bytes)
• Unique for each network interface card made
  (hopefully)

29
Ethernet Addressing

• This computer MAC 00-02-2D-6D-CD-9B (base 16)
• In binary 00000000-00000010-00101101-01101101-110
  01101-10011011

30
Ethernet Frames

• Ethernet divides data to be transmitted into
  frames
• Ethernet frame has six parts
• Preamble (8 bytes)
• Destination MAC address (6 bytes)
• Source MAC address (6 bytes)
• Type or length (2 bytes)
• Data (64 1500 bytes) (usually)
• FCS (4 bytes)

31
10Base2, 10Base5

• Good news - no devices needed to control traffic
  on the network
• Bad news no devices available to control
  traffic on the network

32
Ethernet (Star Topology)

• 10BaseT, 100BaseT(X), 1000BaseT, 1000BaseSX,
  1000BaseLX
• Require a device at center of star
• Ethernet hub or Ethernet switch

33
Ethernet Hubs and Switches

• Hub any frames transmitted by a connected
  computer are sent out all ports (to all connected
  computers)
• Switch learns which computers are connected,
  what port they are connected to, and only
  transmits frames out the port that the specific
  receiving computer is connected to

34
Hubs, Switches, Collisions

• Consider a 4-node 10Base2 network, a 4-node
  10BaseT network with a hub, and a 4-node 10BaseT
  network with a switch
• Which network will have the most collisions?
  the least?

35
Hubs, Switches, Collisions

• A network with a hub is a single collision domain
  (bad!)
• A network with a switch has a separate collision
  domain for each port (good!)

36
Ethernet Hubs and Switches

• Hubs single collision domain, single broadcast
  domain
• Switches multiple collision domains, single
  broadcast domain
• Hubs and switches can be used together in a
  network

37
Ethernet Broadcasts

• A broadcast frame has destination address of
  FF-FF-FF-FF-FF-FF (binary all 1s)
• A switch will send broadcast frames out every
  port (except the one on which the frame was
  received)

38
Ethernet Broadcasts

• Broadcasts are sometimes necessary
• Broadcasts are sometimes evil

39
Broadcast Domains

• Example Consider an Ethernet network with an
  8-port switch fully connected How many
  broadcasts domains are in this network? How many
  collision domains are in this network?

40
How Does A Switch Work?

• It records the source MAC address in every frame
  it receives and stores it in the filter table
  with the associated port from which it came
• If a switch receives a frame destined for a MAC
  address that is not in the filter table, the
  switch will send it out every port

41
Real World Show and Tell

• HP Procurve 2848 switch
• Can mix Ethernet standards on one device
• 1000Base-LX or 1000Base-SX
• 1000Base-T/100Base-T/10Base-T autosensing

42
Thats a Wrap on Ethernet(for now)

• Other LAN technologies
• FDDI (Fiber Distributed Data Interface)
• Token Ring
• LocalTalk (Apple)

43
Remember This?(Lets refine it)
Transmission medium
44
Network Layers OSI Model(Open Systems
Interconnection)
Application Layer
Presentation Layer
Session Layer
Transport Layer
Network Layer
Data Link Layer
Physical Layer
45
Why All the Layers?

• Provides a model for how communication should
  take place
• Real world example organization chart in a
  business (president, VP, mid-managers,
  low-managers, entry-level staff)

46
Why All the Layers?

• Software developers only have to be concerned
  with a particular layers functions
• Allows many companies (vendors) to develop
  software that will work together
• Allows various types of network hardware and
  software to communicate
• Changes in one layer dont cause problems in
  other layers

47
Role of Each Layer in OSI Model

• Application Layer provides an interface between
  the application software (e.g. Internet Explorer,
  AIM) and the lower network layers
• Presentation Layer translates data to standard
  format provides encryption and data compression

48
Role of Each Layer in OSI Model

• Session Layer directs traffic
• (will not be emphasized just know that it
  exists and where)

49
Role of Each Layer in OSI Model

• Transport Layer
• takes streams of data from application software
  and upper layers
• converts data stream into segments
• opens communication with receiving computer
• Provides either reliable or unreliable
  communication to receiving computer

50
Role of Each Layer in OSI Model

• Network Layer
• Manages network addresses
• Responsible for transporting data to other
  computers which may not be attached to the local
  area network
• Takes segments from transport layer
• Sends datagrams (or packets) to data link layer

51
Role of Each Layer in OSI Model

• Data Link and Physical Layers
• This is where Ethernet exists
• Data link layer takes datagrams from network
  layer and builds frames

Preamble (8 bytes)
Data (64 up to 1500 bytes)
Destination MAC Address (6 bytes)
Source MAC Address (6 bytes)
Length (2 bytes)
FCS (4 bytes)
52
Data Encapsulation Through Layers

• Information from layer above is encapsulated (has
  a header and error detection information added)
• Corresponding layer on receiving computer uses
  and then removes the header and error detection
  data (if any)
• More on this later

53
TCP/IP

• Transport and Network Layers Protocols
• TCP Transmission Control Protocol
• Operates at the transport layer (layer 4)
• IP Internet Protocol
• Operates at the network layer (layer 3)

54
TCP/IP

• Developed by Department of Defense in 1960s
• Wanted to connect mainframe and supercomputers in
  different parts of the country

55
TCP/IP

• Wanted the network to not have a single point of
  failure
• End node verification
• Dynamic routing

56
TCP/IP

• This network was called ARPAnet (Advanced
  Research Projects Agency)
• NSF took the design and used it to connect
  research centers and universities
• NSFs network became known as the Internet (Al
  Gore??)

57
Features of TCP/IP

• Logical addressing
• Ethernet cant get us very far!
• Routing (new network device for us!)
• Routers connect networks together
• Data addressed to the local network doesnt go
  through the router

58
IP Addresses

• 32 bit (4 bytes)
• Usually displayed in base 10 notation
• Example 12.146.244.182
• Unique to each computer (but user controllable)

59
IP Addresses

• Network portion
• Host (or computer portion)
• Telephone number analogy
• Subnet mask (netmask) determines boundary

60
Example IP Address

• Example IP Address 206.74.226.4
• binary (base 2) equivalent 11001110.01001010.11
  100010.00000100
• Netmask 255.255.255.0
• binary equivalent 11111111.11111111.11111111.00
  000000
• - The 1s (the on bits) indicate the network
  portion, the 0s represent the host (or
  computer) portion

61
IP Address - Network

• An address with all zeros in host portion is
  generally referred to as the network address.
• Example
• 206.74.226.0
• 11001110.01001010.11100010.00000000

62
IP Address - Broadcast

• An address with all ones in host portion is the
  broadcast address.
• Example
• 206.74.226.255
• 11001110.01001010.11100010.11111111

63
A Rule or Two

• A host cannot have the network address.
• A host cannot have the broadcast address.
• (Basically, an IP address assigned to a host
  cant have all ones or all zeros in the host
  portion of the address.)
• 127.0.0.1 is reserved.

64
IP Addresses

• Three main classes of IP addresses
• Class A
• Class B
• Class C

65
IP Addresses Class A

• Class A
• Intended for the networks with very large number
  of nodes
• First byte of address (first octet) is network
  portion (i.e. netmask 255.0.0.0)
• First bit of first byte of address must be 0
  (binary)
• What is the range of network addresses?
• How many networks?
• How many hosts?

66
IP Addresses - Class B

• Class B
• Intended medium-sized networks
• First two bits of first byte of address must be
  10 (binary)
• First two bytes of address (first two octets) are
  network portion (i.e. netmask 255.255.0.0)
• What is the range of network addresses?
• How many networks?
• How many hosts?

67
IP Addresses - Class C

• Class C
• Intended for smaller networks
• First three bits of first byte of address must be
  110 (binary)
• First three bytes of address (first three octets)
  are network portion (i.e. netmask
  255.255.255.0)
• What is the range of network addresses?
• How many networks?
• How many hosts?

68
IP Address Classes - Summary

• Class A
• Network address range
• 0.x.x.x 126.x.x.x (127 class A addresses)
• Netmask 255.0.0.0
• Class B
• Network address range
• 128.0.x.x 191.255.x.x (16384 class B addresses)
• Netmask 255.255.0.0
• Class C
• Network address range
• 192.0.0.x 223.255.255.x (2,097,152 class C
  addresses)
• Netmask 255.255.255.0

69
IP Addresses (Class D and E)

• They exist
• Not commonly used
• We will not study them

70
Why Is the Netmask Needed?

• If we can look at the first octet in the address
  and tell which class the address is in, why do we
  need to specify the netmask?
• Answer The netmask can be varied to allow
  subnetting, more later

71
Review the Big Picture

• Application software
• Network layers (OSI model)
• Application, Presentation, Session -gt upper
  layers
• Transport layer (TCP is the transport layer
  protocol we are studying)
• Network layer protocol (IP)
• Data Link and Physical layers (Ethernet)

72
Review the Big Picture

• Upper layers produce data stream
• TCP (transport layer protocol)
• takes data
• produces segments
• sends segments to network layer protocol
• IP (network layer protocol takes segments)
• Constructs a packet
• puts segment into data field in packet
• adds IP header (with source and destination IP
  addresses and other info)
• sends packet down to data link layer
• Ethernet (data link layer)
• Constructs a frame
• puts IP Packet into data field in frame
• adds header and FCS fields to frame
• sends frame to physical layer (network interface
  card)
• Physical layer sends the frame onto the medium
  (the wire) as series of bits in the form of
  electrical signals

73
IP Packet (a.k.a. IP datagram)

• Version
• IP version number
• 4 bits
• Header length
• 4 bits
• Priority and type of service
• 8 bits
• Total length
• Length of header and data combined (entire
  packet)
• 16 bits
• Indentifier
• Like a serial number for the packet
• 16 bits

74
IP Packet

• Flags
• Indicates fragmentation
• 3 bits
• Fragmentation
• If packet is too large for frame, provides info
  to help reassemble packet on other end
• 13 bits
• Time To Live
• Expiration time
• 8 bits
• Protocol
• Transport layer info (port number and protocol)
• 8 bits
• Header checksum
• For error detection within IP packet
• 16 bits

75
IP Packet

• Source IP address
• 32 bits (of course!)
• Destination IP address
• 32 bits
• Options
• Used for testing, debugging, etc.
• 0 bits or 32 bits
• Data
• The payload - contains the data from/to the
  transport layer (usually the TCP segment)
• Varies in length

76
IP Packet

• Most important things to remember
• Contains source and destination IP addresses
• Contains TCP port info
• Contains data

77
Examining Incoming Data

• Examine FCS field in frame
• Examine destination MAC address in frame
• Examine header checksum in IP packet
• Examine destination IP address in packet
• If all these pass
• Send data (TCP segment) to TCP for further
  processing

78
Subnetting Example

• Suppose you have a small office network with only
  5 computers/network devices (5 hosts). Assigning
  a class C license to you organization would be
  wasteful of the precious IP addresses.

79
Subnet Example

• Your ISP could assign you an network IP address
  like this
• Network address 220.178.12.144
• Binary 11011100.10110010.00001100.10010000
• Netmask 255.255.255.240
• Binary 11111111.11111111.11111111.11110000
• Broadcast 220.178.12.?
• Binary ?
• How many hosts can be on this IP subnet?

80
What good is IP subnetting?

• Conserves addresses
• Allows a large network to be broken up into
  smaller networks to increase efficiency
• Reduce the broadcasts that hosts receive
• Problems can be contained (broadcast storms)
• Allow network bandwidth to be controlled

81
How do we subnet?

• Router
• Connected to two or more subnetworks
• Forwards packets based on destination IP address
• Each network interface on a router will have an
  IP address assigned to it that is part of the IP
  subnet

82
Review Switching, Broadcasts, Collisions

• Hubs repeat everything
• Switches forward frames based on destination MAC
  (Ethernet) address
• Switches always forward broadcasts
• Every switch port is a collision domain

83
Back to Routing

• Routers
• Do NOT forward Ethernet broadcasts
• Do forward IP packets based on destination IP
  address
• Forward a packet to the network in which the
  destination IP address resides

84
Routing Example 1

• Consider Computer A and Computer B directly
  connected via Ethernet cable (Wow, you can do
  that?)
• Computer A sends data to Computer B
• What happens?

85
Back to Routing Example 1 (contd)

• Computer A 220.178.12.42
• Computer B 220.178.12.43
• The netmask is 255.255.255.0
• (Remember all hosts on a common subnet must have
  a common netmask and network address!)
• What is the network address for this small
  network?

86
Back to Routing Example 1 (contd)

• In Computer A
• Data comes from app s/w and upper layers
• TCP creates a segment, passes down to IP
• IP builds packet with destination IP address,
  source IP address, data (TCP segment), and other
  header fields
• IP determines if the destination IP address is on
  the same subnet as the source IP address (why?
  more on this in routing example 2)
• IP passes packet down to data link layer for
  frame creation but wait!

87
Back to Routing Example 1 (contd)

• The Ethernet frame must have a destination MAC
  address, right?
• No data can be passed from A to B on an Ethernet
  network without a destination MAC address
• What gives??

88
Address Resolution Protocol (ARP) to the rescue

• If IP has a packet to send, it must inform the
  data link layer (Ethernet) of the destination MAC
  address
• ARP serves as IPs detective
• IP uses ARP to find the MAC address that
  corresponds to a particular IP address

89
Address Resolution Protocol (ARP)

• ARP sends out an Ethernet broadcast frame
  (destination address is all 1 in binary or all
  FF in hexadecimal)
• The broadcast frame basically asks, Would the
  host with this IP address please respond to me
  with your MAC address?
• All hosts on the subnet will process the frame
  only the particular host with the destination IP
  address will respond

90
Back to Routing Example 1 (contd)

• After ARPing for the MAC address, IP sends the
  packet down to the data link layer along with the
  destination MAC address
• Data link layer builds the frame
• Passes to physical layer for transmission as
  series of bits yada yada yada

91

• Lets do it again

92
Routing Example 2

• Router in between Computer A and Computer B
• 220.178.12.0, netmask 255.255.255.0
• 220.178.13.0, netmask 255.255.255.0
• Assume router interfaces have following IP
  addresses/netmasks
• E0 220.178.12.1 / 255.255.255.0
• E1 220.178.13.1 / 255.255.255.0

93
Routing Example 2 (contd)

• Computer A sending data to Computer B
• Computer A
• IP address 220.178.12.34
• Netmask 255.255.255.0
• Computer B
• IP address 220.178.13.147
• Netmask 255.255.255.0
• Question What are the network and broadcast
  addresses for the two subnets in this example?

94
Routing Example 2 (contd)

• In Computer A
• Data comes from app s/w and upper layers
• TCP creates a segment, passes it down to IP
• IP builds packet with destination IP address,
  source IP address, data (TCP segment), and other
  header fields
• IP determines if the destination IP address is on
  the same subnet as the source IP address
• If destination is on the same subnet, then ARP
  for the MAC address of computer with destination
  IP address

95
Routing Example 2 (contd)

• But wait! In this example, Computer B is NOT on
  the same subnet with Computer A
• Will ARP work? Remember that the router does
  not forward Ethernet broadcasts and ARP uses an
  Ethernet broadcast

96
Routing Example 2 (contd)

• Computer A must know IP address of default
  gateway for its subnet
• The default gateway is the IP address of the
  router interface on that subnet

97
Routing Example 2 (contd)

• Computer A
• ARPs for the MAC address of the default gateway
  (the router)
• Router responds with MAC address for its Ethernet
  interface on that subnet (E0)
• Computer A sends Ethernet frame to router
  (containing the IP packet with the original
  source and destination address)

98
Routing Example 2 (contd)

• Router
• Sees the frame is for him
• The routers data link layer passes the IP packet
  up
• The IP layer on the router examines the IP
  destination address
• The router sees that the destination is on the
  same subnet with interface E1
• ARPs for MAC address of destination computer
  (Computer B) Computer B responds
• Router builds a frame with recipients real MAC
  address as destination and original IP packet
  payload
• Sends the frame down to physical layer for
  transmission

99
IP Addressing/Subnetting Review

• Example
• IP Network Address 196.24.44.80
• Subnet Mask (netmask) 255.255.255.248
• What is the range of host addresses?
• What is the broadcast address?

100
IP Addressing/Subnetting Review

• Network Address
• 11000100.00011000.00101100.01010000
• Netmask
• 11111111.11111111.11111111.11111000
• Broadcast
• 11000100.00011000.00101100.01010111
• First host is network address 1
• 11000100.00011000.00101100.01010001
• Last host is broadcast 1
• 11000100.00011000.00101100.01010110

101
IP Addressing/Subnetting Review

• First host is network address 1
• 11000100.00011000.00101100.01010001
• 196.24.44.81
• Last host is broadcast 1
• 11000100.00011000.00101100.01010110
• 196.24.44.86
• Range of host addresses on this subnet
• 196.24.44.81 -gt 196.24.44.86

102
Routing Example 3
E0 206.113.116.169
C
Router A
Switch
E2 221.19.10.1
E1 220.178.13.2
E0 220178.13.1
B
Router B
A
Switch
Switch
E2 220.178.17.161
E1 220.178.12.145
103
Routing Example 3 (contd)

• Computer A to send IP packet to Computer C

104
Routing Example 3 (contd)

• Computer A
• ARPs for the MAC address of the default gateway
  (router A)
• Router A responds with MAC address for its
  Ethernet interface on that subnet (E1)
• Computer A sends Ethernet frame to router A
  (containing the IP packet with the original
  source and destination address)

105
Routing Example 3 (contd)

• Router A
• Sees that the frame is for him (destination MAC
  address)
• The routers data link layer passes the IP packet
  up
• The IP layer on the router examines the IP
  destination address
• The router sees that the destination is NOT on
  any subnet to which he is connected
• Router A discards (drops) the packet
• The End

106
Routing Example 3 (contd)

• How can this be made to work?
• Solution 1 Configure a default route on router
  A
• Similar to default gateway on computers
• Default route is the IP address on a local subnet
  to which all packets destined for foreign IP
  addresses are forwarded

107
Routing Example 3 (contd)

• Router A would have in its configuration
• gt Default route 220.178.13.2
• (The IP address for E1 on router B)

108
Routing Example 3 (contd)

• Now, what will router A do?
• Sees that the frame is for him (destination MAC
  address)
• The routers data link layer passes the IP packet
  up
• The IP layer on the router examines the IP
  destination address
• The router sees that the destination is NOT on
  any subnet to which he is connected
• Router A ARPs for MAC address corresponding to
  default route (gateway) address
• Gets a reply, sends frame to E1 on router B

109
Routing Example 3 (contd)

• Router B
• Sees that the frame is for him
• Unpacks the frame and sends data up to IP
• IP sees that the destination IP address is on the
  same subnet with interface E2
• ARPs for MAC address of destination computer
  (Computer B) Computer B responds
• Router builds a frame with recipients real MAC
  address as destination and original IP packet
  payload
• Sends the frame down to physical layer for
  transmission

110
Routing Example 3 (contd)

• Solution 2 Configure a static route on router A
• Simply tells router A to send all packets
  destined for a particular foreign network to a
  specific local IP address
• In this example, configure router A with
  following command
• gt 221.19.10.0 via 220.178.13.2

111
Routing Example 3 (contd)

• What routing configuration does router A need to
  allow hosts full access to LAN hosts and the
  Internet?
• What about router B?

112
IP Routing Summary

• Default routing the IP address on a local
  subnet to which all packets destined for foreign
  IP addresses are forwarded
• Static routing IP addresses on a local subnet
  to which all packets destined for particular
  foreign IP addresses are forwarded

113
IP Routing Summary

• Routing information in a router is contained in
  the routing table
• Example of a routing table
• 192.168.50.0 255.255.255.0 connected to E0
• 192.168.51.0 255.255.255.0 connected to E1
• 192.168.40.0 255.255.255.0 via 192.168.50.1
• 192.168.30.0 255.255.255.0 via 192.168.51.1
• 0.0.0.0 via 192.168.51.1

114
Dynamic Routing

• Routers educate each other about networks to
  which they are connected
• A protocol for exchanging route information among
  routers is called a routing protocol

115
Dynamic Routing

• The most famous routing protocol in use is called
  Routing Information Protocol (RIP, very
  creative, huh?)
• In RIP, a router will report all of the networks
  to which it is connected and also the number of
  hops (or routers) between it and the particular
  networks
• Also propagates RIP info it has received

116
Dynamic Routing

• Upon receipt of RIP info from a neighboring
  router, all hop counts are incremented by 1 and
  the info is placed into the routing table

117
Dynamic Routing

• Example 3 routers (next slide)

118
Dynamic Routing Example (RIP)
E0 17.14.210.32
E0 206.79.211.44
E1 177.100.48.2
E1 192.168.21.32
A
E3 192.168.21.33
E1 177.100.48.3
B
C
E2 12.34.25.147
E2 186.18.90.97
(assume all netmasks are 255.255.255.0)
119
Dynamic Routing Example (RIP)

• Router A reports to router B

Network Hops
17.14.210.0 0
12.34.25.0 0
120
Dynamic Routing Example (RIP)

• Why is the hop count important?
• A router might receive route information for a
  particular network from two directions
• When this happens, the router will only keep the
  route with the smallest hop count (closest path
  to the network)

121
Dynamic Routing Example (RIP)

• Router B will add these entries in routing table
• 17.12.210.0 via 192.168.21.32
• 12.34.25.0 via 192.168.21.32

122
Dynamic Routing Example (RIP)

• Router B reports to router C

Network Hops
17.14.210.0 1
12.34.25.0 1
206.79.211.0 0
192.168.21.0 0
186.18.90.97 0

• What will router C add to its routing table?

123
Dynamic Routing

• RIP is being gradually being replaced by newer
  more efficient routing protocols
• Open Shortest Path First (OSPF) is becoming
  prevalent

124
Layers Again

• Upper layers (s/w application) -gt Transport Layer
• Transport -gt Network -gt Data Link -gt Physical
• Layers talk to their counterparts
• At what layers do routers operate?
• How does the requirement for end node
  verification fit in?

125
Layers Again

• Transport layer is the first layer in which the
  end nodes really talk to each other
• Transport layer is where end node verification
  takes place

126
Transport Layer (Layer 4)

• An interface for network applications provides
  a way for application software to access the
  network. The designers wanted a way to send data
  not just to a particular computer, but to a
  particular network application running on the
  destination computer

127
Transport Layer (Layer 4)

• Provide multiplexing/demultiplexing the
  transport layer must be capable of simultaneously
  supporting several network applications and
  directing data to the network layer
• Provide mechanism for one network application to
  maintain connections with more than one computer

128
Transport Layer (Layer 4)

• Error checking
• Similar to network and data link layer error
  checking (nobody dont trust nobody)
• Flow control
• One computer doesnt allow the other computer to
  overwhelm it with data
• Verification
• Making sure all the data got delivered

129
Transport Layer (Layer 4)

• Two transport layer protocols
• Transport Control Protocol (TCP)
• provides extensive error checking and flow
  control to ensure successful delivery of data
• It is connection-oriented
• User Datagram Protocol (UDP)
• Provides very basic error checking
• Reliability sacrificed for speed and efficiency
• It is connectionless

130
Transport Layer (Layer 4)

• Oversimplified example two humans in
  connection-oriented conversation
• Bill Hello Larry. Are you listening? I have
  something to say.
• Larry Yes, Im listening Bill.
• Bill There is
• Larry Yes, I understand.
• Bill a baseball game
• Larry Yes, I understand.
• Bill on Saturday.
• Larry Yes, I understand.
• Bill Thats all I have to say.
• Larry Ok, Ill stop listening to you.
• Bill Ok, Ill stop talking to you.

131
Transport Layer (Layer 4)

• Oversimplified example two humans in
  connectionless conversation
• Bill Larry, there is a baseball game on Saturday.

132
TCP and UDP Ports

• Network software applications access the
  transport layer protocols through a port
• Ports are numbered only one software
  application can use one port number at a time
• The ports are not real, hardware ports they are
  software ports

133
TCP Port Example

• Example Computer A wants to download a web page
  from computer B
• Computer Bs web server software is accepting
  connections on TCP port 80
• Computer A will pick an unused port number at
  random and open a connection to computer B on its
  port 80

134
TCP Port Example
Computer B (web server)
Web server software
Network Layer (IP)
Data Link Layer (Ethernet)
Computer A
135
TCP Port Example

• The web server software on B has notified TCP
  that it wishes to accept connections on port 80
  (passive mode)
• The browser software on computer A then asks TCP
  (on computer A) to open a connection to port 80
  on computer B
• Computer A will use a random port number not in
  use already

136
TCP Port Example
Computer B (web server)
137
Well Known TCP Ports

• 20, 21 FTP
• File Transfer Protocol
• 23 Telnet
• Terminal emulation interface
• 25 SMTP
• Simple Mail Transfer Protocol
• 53 DNS
• Domain Name Service
• 80 HTTP
• Hypertext Transfer Protocol (the web)
• 110 POP3
• Post Office Protocol (checking email)

138
TCP Segment

• Source port (16 bits)
• Port number used by transmitting host (max 65534)
• Destination port (16 bits)
• Port number used by receiving host (max 65534)
• Sequence number (32 bits)
• Number corresponding to first byte of data it
  will send
• Acknowledgement number (32 bits)
• The next sequence number that the receiver is
  expecting
• Data offset (4 bits)
• Length of the header (integer multiple of 32 bits)

139
TCP Segment

• Reserved (6 bits)
• All zeroes, all the time
• Control flags (1 bit each)
• URG
• ACK
• PSH
• RST
• SYN
• FIN

140
TCP Segment

• Window (16 bits)
• The next sequence number that the transmitting
  computer is free to send without further
  acknowledgement
• Checksum (16 bits)
• Error correction (similar to lower layers)
• Urgent pointer (16 bits)
• Basically, a sequence number at which some urgent
  data will begin
• Options (variable length)
• Usually either 0 bits or 32 bits
• Padding (variable)
• Extra zero bits to make sure the header is
  integer multiple of 32 bits
• Data (variable length)

141
TCP Segment Most Important Fields

• Source port
• Destination port
• Sequence number
• Acknowledgement number
• Window
• Data

142
Establishing a TCP Connection(Three-Way
Handshake)

• From previous example
• 1) Computer A sends a segment to computer B
  requesting synchronization basically a
  request to open a connection (session)
• This segment also contains As initial sequence
  number
• 2) Computer B sends a segment back that
  acknowledges the synchronization and contains
  its initial sequence number

143
Establishing a TCP Connection(Three-Way
Handshake)(contd)

• 3) Computer A acknowledges receipt of computer
  Bs initial sequence number

144
TCP Flow Control

• The receiving computer, in order to prevent the
  transmitting computer from overwhelming it with
  data, uses the Window field is used to define how
  many bytes of data the transmitting computer can
  send before an acknowledgement

145
TCP Flow Control (illustration)(A sending data
to B)
B
A
TCP Segments
3 bytes of data (1,2,3)
Acknowledge 4, window 5
5 bytes of data (4,5,6,7,8)
Acknowledge 6, window 2
2 bytes of data (6,7)
146
TCP Flow Control

• Its possible that segments will arrive at the
  receiving computer in the wrong order (order
  different than transmitted)
• This may be due to a router going down and the
  route between the two computers being changed
  (dynamic routing)
• TCP can put segments back in the correct order
  before giving data to application software

147
UDP Flow Control (?)

• None!

A
B
8 bytes of data (1-8)
148
UDP Datagram

• Source port (16 bits)
• Port number used by transmitting host (max 65534)
• Destination port (16 bits)
• Port number used by receiving host (max 65534)
• Length (16 bits)
• Length of the entire datagram
• Checksum (16 bits)
• Error detection
• Data (varies)

149
Firewalls

• Definition Hardware and/or software designed to
  a provide security for a network or a particular
  computer
• Can control access based on
• network layer (layer 3)
• transport layer (layer 4)
• Application s/w

150
Typical Firewall Configuration(as a standalone
network device)

• Two network interfaces
• Inside interface (trusted)
• Usually connected to internal corporate/office/cam
  pus network
• Outside interface (not trusted)
• Usually connected to Internet (via Internet
  service provider

151
Typical Firewall Configuration

• Hosts on the network on the inside interface
  usually have unrestricted ability to open TCP
  connections (and send UDP datagrams) to hosts on
  outside
• Exceptions can occur
• Disallow access to certain web sites
• Disallow email to be sent through external mail
  servers (virus/worm control)

152
Typical Firewall Configuration

• Hosts on the outside interface (the rest of the
  Internet) usually have no ability to open TCP
  connections or send un-requested UDP datagrams to
  hosts on inside network
• Exceptions
• Allow external hosts to access a web server on
  port 80 (HTTP port)
• Allow external hosts access to a mail server on
  port 25 (SMTP) for delivering email

153
Typical Firewall Configuration (NAT)

• Most firewalls capable of Network Address
  Translation (NAT)
• Allows for a private IP addressing scheme on the
  inside network
• When inside hosts need to communicate with hosts
  outside, the firewall translates the inside
  (private) IP address to a real IP in outgoing IP
  packets

154
Typical Firewall Configuration (NAT)

• For an IP packet coming back from an outside
  host, the firewall will translate the destination
  IP address back to the particular hosts private
  (inside) address
• When the session is over, the outside IP address
  can be recycled to be used for another inside
  host

155
Typical Firewall Configuration (NAT)

• Advantage can allow a large number of hosts on
  inside network to share a relatively small number
  of real IP addresses for Internet use
• Very important for home networks with more than
  one computer (together with PATmore later)

156
Typical Firewall Configuration (NAT)

• Network address translations can be static so
  that an inside host will always have a particular
  outside (real) IP address
• This is necessary for web servers, email servers,
  DNS servers, or any computer that may need to
  allow incoming connection requests