This slide deck was used for a 2-day short course at IIT Gandhinagar in Spring 2015. Being a 2-day course, it focuses more on a qualitative description of how we access the Internet.

1. INTRODUCTION TO
COMPUTER NETWORKING
Amit Saha
IIT Gandhinagar
13th, 14th March 2015

2. Ground Rules
• There are no “silly” questions. Focus on “why”
• Slides with have fundamental concepts
• Search on the web for anything you don’t understand
• Almost everything is explained reasonably well on the web
• You can send questions to amsaha+iitgn15s@gmail.com

3. Books – not mandatory
• Data and Computer Communications
- William Stallings
• Internetworking with TCP/IP Volume 1 Principles, Protocols,
and Architecture
- Douglas E. Comer

4. What is Networking?
• networking – the exchange of information or services among
individuals, groups, or institutions
• computer networking – the exchange of information or services
among computers

5. A Brief History
• http://visual.ly/brief-history-computer-network-technology

6. Schematic of a Computer Network

7. The first hop – wired

8. The first hop – wireless

9. Wireless First Hop - Possibilities
• WLAN/WiFi
• Cellular

10. Layering – Lets Break up the Problem
TheoreticalPractical

11. Wireless First Hop – WiFi
• IEEE 802.11 (a/b/g/n/…) standards
• Infrastructure mode (common) or Ad hoc mode
Ethernet Token Ring
New 802.11n

12. Connecting to a WiFi Access Point
• Scanning
• Joining
• Authentication
• Association
• The station feels as if it is “plugged” into a wired network

13. Communicating with a WiFi AP
• Wireless is a half duplex channel – either speak or listen
• Signal strength is inversely proportion to square of distance
• Carrier Sense Multiple Access / Collision Avoidance (CSMA/CA)
• Fancy name for how humans talk 
• This is generally true of all wireless systems (including humans)

14. WiFi Equipments

15. Wireless First Hop – Cellular
• Not covered in this course

16. Wired First Hop - Ethernet
• IEEE 802.3 standard
• If two ends of the Ethernet cable (e.g., RJ45) is electrically up,
connection is up.
• Some authentication steps are usually there but we will skip those

17. CSMA/CD
• Carrier Sense Multiple Access / Collision Detect

18. Ethernet Equipments

19. Now what?

20. Getting an IP address
• DHCP – DHCP clients get IP addresses and networking
parameters from DHCP server
• Based on BOOTP
• DHCPv6 (for IPv6) is also available

21. DHCP Packets
• Discovery
• Offer (possibly multiple)
• Request
• Acknowledge
• Renew
• Release
Why do we need an IP address?

22. IP Packet Format

23. IPv4 Header Format

24. UDP Header Format

25. 5-Tuple/Flow

26. Got an IP! Hurray!
• But now what? How do I “go to” some website?
• IP is just an identity
• Still need to “route” to somewhere

27. DNS – Domain Name Service
• We almost always try to reach a human readable name
• But machines route based on IP addresses
• There must be a mapping from names to IP addresses
• But who is going to do this translation?

28. Default Gateway
• If host does not know how to route, who does it ask?
• There must be a “goto” guy in the network

29. Routing table – Host
ubuntu@ubuntu-VirtualBox:~/Projects/OpenStack/New/cinder$ route
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
default 10.0.2.2 0.0.0.0 UG 0 0 0 eth0
10.0.2.0 * 255.255.255.0 U 1 0 0 eth0
• Routing table becomes very interesting in a core router
• 50,000 to 1,00,000 routes (entries like the above table)
• We will get to it later

30. Who could …
• Be a host’s Domain Name Server
• Be a host’s Default Gateway

31. Are we ready to send?

32. ARP – Address Resolution Protocol
• What is the MAC address for a given IP?
• ARP resolves network layer address to link layer address
• Remember network layer is the same as routing layer. This is also known
as Layer 3(L3)
• Link Layer is Layer 2 (L2)

33. Ethernet – IEEE 802.3b
• Broadcast medium
• These days mostly switched Ethernet (point-to-point)
• These days mostly Ethernet II
• No LLC

34. Finally!
• We can now send

35. Ethernet Header

36. An Example
• Wireshark – Download it, install it and then play around with it
• Instructions are there on the web itself
• Will give you a lot of detail about packets and layering

37. DNS – Domain Name Service
• Hierarchical way of managing name space
• Managed by Internet Assigned Numbers Authority (IANA)
• http://www.iana.org/domains/root/db

39. DNS – Root Servers
• <a-m>.root-servers.net – 13 hostnames, around 500 instances
• Use anycast to reach “nearest” instance – BGP supports this
• 12 of the 13 root servers A-M exist in multiple locations
• 11 on multiple continents
• Root server ‘h’ exists in two U.S. locations.
• Root server ‘b’ exists in a single location in the Los Angeles Area
• http://www.root-servers.org/

40. Subnetting
• Routing to each individual hosts does not scale
• Addresses grouped into different classes

41. CIDR
• Any “class system” is bad 
• IPv4 was running out of unique IP addresses
• Hence IPv6 but…
• Classless Inter Domain Routing
• Network portion can be of any size
• Addresses need to specify subnet mask as well

42. Longest Prefix Match

43. NAT – Network Address Translation

44. NAT – Send

45. NAT – Receive

46. Routing Hierarchy
• Internet is divided into Autonomous Systems (AS)
• Managed by a single administrator
• 16 bit AS Number (ASN)

47. Autonomous Systems

48. Intra-Domain Routing
• Typically uses UDP
• Distance Vector
• RIP – Bellman – Ford Algorithm
• Link State
• OSPF – Dijkstra’s shortest path
• ISIS – Dijkstra’s shortest path
• Note: The fundamental concept here is Distance Vector and
Link State as two types of algorithms, not that they are used for
intra-domain routing

49. Inter-Domain Routing
• Border Gateway Protocol (BGP) – BGP version 4
• Path – vector routing
• Uses TCP
• External BGP – eBGP
• Internal BGP – iBGP
• Not to be confused with intra-domain routing (though iBGP could do that )
• Used to send information from one “side” of network to another
• Policy engineering

50. Inter – Domain Routing

51. iBGP and eBGP

52. Some Numbers
• http://www.cidr-report.org/as2.0/

54. Transport Layer
• Provides end-to-end connection
• On top of the routing layer
• Two types of transmissions
• Unreliable – User Datagram Protocol (UDP)
• Reliable – Transmission Control Protocol (TCP)

55. UDP
• Best effort, connection-less
• Just send as fast as possible
• When is it okay?
• Phone calls, gaming, etc.
• Simple stuff like DNS

56. TCP
• Connection oriented
• Setup/teardown connection
• Sends stream of bytes, not messages
• Reliable and in-order delivery
• Flow Control
• Congestion Control
• Analogous to how humans communicate

57. Basics of Reliable Connection
• How do you make a transmission reliable?
• What do you need to protect against?

58. Basics of Reliable Connection
Data
Ack
Sender Receiver
time

59. Speed it Up

60. TCP – Connection Setup
• Three-Way Handshake

61. TCP Sliding Window
• http://www2.rad.com/networks/2004/sliding_window/

62. TCP – Being a Good Citizen
• Congestion Control
• MaxWindow = min(Congestion Window, Receiver Window)
• Congestion Window start at 1
• Original TCP: Additive Increase, Multiplicative Decrease

63. Congestion Avoidance Behavior
Time
Congestion
Window
Packet loss
+ Timeout
Grabbing
back
Bandwidth
Cut
Congestion
Window
and Rate

64. Evolution of TCP
1975 1980 1985 1990
1982
TCP & IP
RFC 793 & 791
1974
TCP described by
Vint Cerf and Bob Kahn
In IEEE Trans Comm
1983
BSD Unix 4.2
supports TCP/IP
1984
Nagel’s algorithm
to reduce overhead
of small packets;
predicts congestion
collapse
1987
Karn’s algorithm
to better estimate
round-trip time
1986
Congestion collapse
observed
1988
Van Jacobson’s algorithms
congestion avoidance and
congestion control
(most implemented in
4.3BSD Tahoe)
1990
4.3BSD Reno
fast retransmit
delayed ACK’s
1975
Three-way handshake
Raymond Tomlinson
In SIGCOMM 75
Taken from Aditya Akella’s slides, Dept. of Computer Science, University of Wisconsin - Madison

65. TCP Through the 1990s
1993 1994 1996
1994
ECN
(Floyd)
Explicit
Congestion
Notification
1993
TCP Vegas
(Brakmo et al)
real congestion
avoidance
1994
T/TCP
(Braden)
Transaction
TCP
1996
SACK TCP
(Floyd et al)
Selective
Acknowledgement
1996
Hoe
Improving TCP
startup
1996
FACK TCP
(Mathis et al)
extension to SACK
Taken from Aditya Akella’s slides, Dept. of Computer Science, University of Wisconsin - Madison

66. TCP Header

67. Ethernet
• Original Ethernet used hub – shared ethernet
• These days – switched ethernet

68. Switched Ethernet
• CSMA / CD
• What is collision now? How is it handled?

69. Switching Types
• Store and forward
• Cut through

70. Switched Ethernet – Arbitrary Topology
• Can cause loops
• Unlike IP header, no TTL
to prevent looping

71. STP – Spanning Tree Protocol
• Creates a “tree” that “spans” entire graph
• Distributed implementation

72. STP – Root Bridge

73. STP – Root Ports

74. STP – Designated Ports

75. STP – Blocked Ports

76. STP – Link Failure

77. Virtual LAN

78. VLAN

79. IEEE 802.1q

80. IEEE 802.1q

81. STP Variants
• Rapid STP – Converges faster than STP
• VLAN STP – Each VLAN can have a different STP tree

82. Data Centers
• Cloud computing – Making computing elastic
• Software Defined Networking (SDN)
• Universities are using data centers for compute resources
• Pay-as-you-go pricing model
• Power and Cooling most important 

83. Data Centers – Network Properties
• Large number of ports
• Large number of L2 end points
• Multiple VMs per port
• Relatively fewer L3 end points
• Prime importance - manageability

86. Example Equipment – Pictures/Prices
• Go to websites such as www.cisco.com and www.juniper.net
and look at their product portfolio
• www.cisco.com has interactive 3D models
• Search for prices on the web

87. Industry Expectations

88. Industry Expectations
• Average loss $5600/minute  $300K/hour
• Zero planned downtime is fast becoming the norm
• Can you think of anything similar?

89. The Future of Computer Networks
• Manageability
• Monitoring
• Trouble shooting
• Automated Healing
• More bandwidth
• Energy efficiency
• Faster way of implementing something –
algorithms/algorithmics
• OpenSource based – cheaper and “better”
• Broadband access will become a fundamental right