Transcript internetworks - UNC School of Information and Library Science

How the Internet Works?
( TCP/IP, DNS… )
How computers send data?
Protocol
Connection
method
Channel
Address
2
Communication Channel
what kind of media?
 Telephone line (Twist pair, Optical fiber)
 Modulator-Demodulator (Modem)
 Digital Subscriber Line (DSL)
 Cable modem
 Satellite, Microwave
 Wireless connection (IR, RF)
3
Different medium
 Twisted pair
 RJ45 connector
 Coaxial cable
 BNC connector
 Fiber optic cable
 Different cabling length
 Optic > coaxial > Twisted pair
4
Use Modem
Digital data - 10101010101
Telephone line
Analog data

Modulation / De-modulation

E.g. CU Dialup Pool (56kbps max.)
5
Use Digital Subscriber Line
(DSL)
Traditional phone line
ADSL modem
Why ADSL?
6
Why ADSL?
 Asymmetric DSL
 A slower upstream
(upload) can trade off a
faster downstream
(download) speed.
 128 to 640 kbps (upstream)
 1.5M to 5M bps
(downstream)
7
Use Cable Modems
 Use coaxial cable to carry TV signal and High speed
Internet access
8
Satellite &
Microwave
 Satellite systems
 Use satellites orbiting above the Earth to relay signals
from one part of a WAN to another, cause 0.5 to 5 sec
delays.
 Microwave
 Costly to install but cheaper than satellite.
 It is very useful for connecting networks that are
separated by a barrier, such as a highway or a lake.
 Requires line of sight.
9
Using A Computer To
Interconnect Networks
 Special-purpose
computers are used to
interconnect networks.
 Using standard hardware
(CPU, memory, and
network interfaces)
 Running special-purpose
software
10
Network Interface Card
NIC
 Physically connects a computer to the
transmission medium on a network.
11
Hardware/Physical/MAC (Media
Access Control) address
 When a NIC is
manufactured, the
card is given a
unique hardware
address.
 It never changes.
12
Routers
 Interconnecting computers are called routers by
using the same protocol.
 Determining where to send packets
Router
13
Hubs
 A network cable connects a
computer via a network card
to a hub.
 Provides a central location.
14
Ports
 A hub
contains
sockets or
ports.
 Some LED
indicates
information
transferred
through the
port.
15
Ethernet Repeater
 A repeater is a device that strengthens and
retransmits signals on a network.
16
Network Architecture
 It refers how information transfers on networks.
 Ethernet
 It the most popular architecture used to build networks.
 Least expensive and easiest to setup
 Token-ring architecture
 It was developed by IBM in 1984.
 They are popular found in large organizations, such as
banks and insurance companies.
 Others: ARCnet, AppleTalk, …
17
Ethernet & Token Ring
 Ethernet, Fast Ethernet, Gigabit Ethernet (transmit
data at 1Gbps).
18
Postal Services
 You have to write a complete address on
the envelope specifying the country, state,
city, district, street, and so on.
 After put the letter put into the mailbox, it
will be delivered (routed) to its destination
in a hierarchical way.
 California Post Office knows the letter is
sent to NC, without concerning the actual
address to be routed within NC.
19
Basic Internet Structure
 The Internet works in a similar way as postal
services.
 Roughly speaking, you may consider states
within this country, as individual networks
connecting to each other.
 The different pieces of the Internet are
connected by a set of computers (Gateways)
 Translates between protocols
20
Postage over the Internet
 When delivering information via the Internet, the
information is split into small units called data
packets (1500 byte each)
 When a packet is sent from California to a
particular host in UNC.




The packet is first delivered to NC,
then is further transmitted to UNC,
then is further transmitted to the appropriate
department,
finally arrived to the specific host.
 The data is reassembled at the destination.
 The data packet is continuously being switched
from the source to destination.
 The Internet is said to be Packet Switching
Network.
21
Packet Switching Network
22
Packet switching example
Figure 16.1 An example
internet with four networks
connected by routers.
Figure 16.2 Cars from two roads merging
onto another road are analogous to packets
from two networks merging onto a third
network.
23
IP address
 Each host in the Internet is assigned to a
specific and unique number for
identification.
 This number is called the IP address of
the specific host.
 This number is divided into 4 parts for
improving the readability.
 The range of each number is between 0
and 255.


E.g. 0.0.0.0
255.255.255.255
 For example, the host “UNC.edu” has
its IP address of “152.2.240.8”
24
Network Number /
Host Number
 IP addresses are split into 2 parts
 A network number + a host number
 For example, 152.2 is the network number of
UNC, 240.8 is the host number of the host
“UNC.edu”
 Network numbers are assigned by a central
authority, the Internet Corporation for
Assigned Names and Numbers (ICANN).
25
Network Classes
 There are 5 classes of IP address
 Class A comprises networks 1.0.0.0 to
127.255.255.255, the network address is in
first quad. It allows roughly 16 million hosts
per network.
 Class B comprises network 128.0.0.0 to
191.255.255.255, the network address is in
the first two quads. It allows for 16,382
networks with up to 64K hosts.
26
Network Classes
 Class C comprises networks 192.0.0.0 to
223.255.255.255, with the network number
contained in the first three quads. It allows about 2
million networks with up to 254 hosts in each
network.
 Class D and E are falling into the range of 224.0.0.0
to 254.0.0.0 which are reserved for multicast address
and for special purpose use.
27
Subnet
 The Internet is structured hierarchically. UNC consists of
many academic departments and administrative bodies.
 IP allows you to subdivide a network into several
subnets. E.g. SILS and Davis Library are two subnets
inside UNC.
 Each subnet is identified by a subnet number.
 E.g. we have a different way to interpret the IP address
152.2.62.223 (the ILS server)
 152.2 refers to the network number of UNC,
 62 refers to the subnet number of Davis Library,
 223 refers to the host number of “afton”.
28
Dynamic Host Configuration
Protocol (DHCP)
 A Protocol for assigning dynamic IP address
to devices on a network.
 It is built on client and server models.
 Server is the machine running DHCPD.
 Client can be any network devices.

Advantage?
Eliminates manual
configuration of network
parameters and utilizes the use
of IP address
29
IP Routing
 The process of transmitting a data packet from the
source to the destination via a series of intermediate
stations is called “Routing”.
 IP routing works as follow:
 Each data packet is labeled with IP address of the
destination host
152.2.90.184 1500 bytes Data here
30
IP Packet
 Includes Header, payload, data
31
Packets Are Not The Same Size
 Packets may be any size up to the maximum.
 Can be as small as a single keystroke
 Can be larger, depending on application
32
IP Routing
 For example, when a packet is routing to
“UNC.edu” (152.2.62.223), the network
number is extracted as 152.2 which is the
network number of UNC. The packet is then
sent to NC and then to UNC.
 Inside UNC, the subnet number is examined
and it is 62 which is the subnet number of
Davis. Then it is sent to Afton.
 Inside Davis, the host number is examined
and it is 223 which is the host number of
Afton. Finally, it is routed to the destination.
33
IP enough for routing?
 A single packet is limited in length, usually 1-1500
bytes.
 Network may lose packets, or damage the data in
transit.
 Packets may arrive out of sequence (different
routing path).
 TCP is used to solve the problems.
34
Protocol and Addressing
 To communicate over the Internet, the
computers must:
 use a common language or a protocol to govern
the exchange of messages.
 have a way to address one another.
 Protocol:
 specifies exact format, order of messages sent
and received among network entities, and
actions taken on message transmission and
receipt.
 Addressing:
 defines where to deliver the messages.
35
Protocol and Addressing
TCP/IP model
36
Protocol and Addressing
 Internet has a large collections of protocols
organized in a layering model.
 Application: enables the user, whether human or
software, to access the network.
 Transport: responsible for source-to-destination
(end-to-end) data transfer.
 Network: responsible for routing packets from
source-to-dest across multiple networks.
 Data link: responsible for data transfer between
neighboring network elements.
 Physical: coordinates the functions required to
transmit a bit stream over a physical medium.
37
Protocol and Addressing

Lower layer adds header to the data from upper layer.
Header includes addressing and other fields.
L5 data
L5 data
38
Protocol Stack - Open Systems
Interconnection (OSI) model
39
Protocol and Addressing

TCP/IP Protocol Suite.
40
What is TCP/IP?
 “TCP would be in charge of the breaking
up the packets and messages then
reassembling them at the destination, and
the IP would be responsible for transmitting
the individual packets. For example: the TCP
protocol would split up the letter and place
it into multiple envelops, while the IP
protocol would be in charge of addressing
the envelop and making sure it arrived at its
proper destination.”
 – from “Where Wizards Stay Up Late”
41
TCP/IP
 A protocol is a collection of rules for formatting,
ordering, and error-checking data sent across a
network.
 In 1974, Vincent Cerf and Robert Kahn developed
the Transmission Control Protocol (TCP) which was
further split into the Internet Protocol (IP) and TCP
in 1978.
 In 1982, DoD adopted TCP/IP as the standard
protocol in the Internet.
 Because the significance of TCP/IP in the history of
the Internet, Cerf and Kahn are considered to be the
Father of the Internet.
42
Transmission Control
Protocol - TCP
 Basic functions
 Decompose a lengthy data into multiple packets for
transmission
 Error detection, ensure validity
 Packet loss?
 No problem, packet retransmission
43
Sequence Numbers
 TCP breaks the information into multiple
packets.
 Each packet is associated with a sequence
number for identification.
152.2.90.184 Number 1 Data here
152.2.90.184 Number 2 Data here
152.2.90.184 Number 3 Data here
 Each packet is individually routed in the
Internet, and arrive in random order.
 The data is reassembled in the correct order
according to the sequence number.
44
Packet Retransmission
 A packet may be lost during the
transmission across the Internet (host down,
link failure, … )
 When the destination host has been waiting
for a particular packet for a certain time
(timeout), it will request the source host to
retransmit the packet.
 There is no need to retransmit all data
packets. Instead, only the missing packet,
which is identified by the sequence number,
needs to be retransmitted.
45
Error detection Checksums
 Transmission errors occur even if a data
packet is received by the destination
successfully.
 How to ensure the data is correctly received?
 A method to detect possible transmission errors.
 At the destination, checksum is recalculated
based on the received data.
 The attached checksum and the newly calculated
checksum are compared. Mismatch means there
is transmission errors occurred.
46
IPv6 (IP version 6)
 Major changes:
 More addresses
 IP address size from 32
bits to 128 bits
 Simplified IP headers
 Reduction of header fields
in IP packet
 Added security features
47
IPX/SPX Protocol
 IPX/SPX makes up the protocol
suite that is used to transfer
information on networks
running the Novell NetWare
operating system.
 Internetwork Packet Exchange
(IPX) - Transfer information
between devices.
 Sequenced Packet Exchange
(SPX) - An extension of the IPX
protocol.
48
NetBEUI Protocol
 NetBIOS Extended User Interface
 Is a network protocol used on small local area
networks.
 A very small and efficient protocol, use little computer
resources.
 NetBIOS
 Network Basic Input/Output System
 Develop by IBM and allow computers to
communicate with each other on a network.
49
Why DNS?
 What is DNS?
 Domain Name System
 IP address is difficult to remember.
 152.2.92.1 is which machine?
 Names are given to each computer on the
Internet for the convenience of human users.
 Besides IP addresses, all internet
applications allow users to use computer
names.
50
Computer names on the
NET
 How does it look like?
 DNS administrators is responsible to name
computers/group in their own subnet.
 Each level of responsibility is called a domain.
 Domains are separated by “dots”
 cse.UNC.edu.NC
 www.intel.com
51
Domain name and Host
name
 Each domain can create or change whatever belongs
to it.
 UNC can create any new domain, cse
 CSE can buy a new computer and name it as – robin,
orchid, any other name.
 Two computers may have the same name if they are
in different domains.
 UNC.edu
 orchid.ie.UNC.edu.NC
52
Domain Name System (DNS)
 Each node in the tree has
 a label - a string with a
maximum of 63
characters.
 a domain name - a
sequence of labels
separated by dots.
53
How does DNS work?
 DNS server = DNS service?
54
IP vs Domain Name
55
Most Common hostname
Figure 18.1 The fifty most common names assigned to computers on the Internet in 2000.
56
Top Level Domains
 The Internet Corporation for Assigned Names and
Numbers (ICANN) define the Top Level Domains.
57
ISO 3166 Country Codes
 Partial list
Full list
58
Windows Internet Naming
Service (WINS)
 Like DNS, WINS
also resolves names
and IP addresses
except that the look
up by WINS is
specific for
Windows
computers.
 Dynamic
 Win2k use DNS to
replace WINS
59
Summary
 What should be considered if you install or update a
network?








Plan: the size, the budget
Determine: the network design
Select: the transmission media
Choose: the network hardware, devices
Install: the network OS, drivers
Configure: server & client computers
Test: the final network
Maintain: the network reliability, admin
60