?Error%20control?%20%20?%20Network%20architecture%20?%20%20?%20Protocols%20?%20%20?%20Transmission%20Efficiency%20and%20Throughput%20?

1
?Error control? ? Network architecture ? ?
Protocols ? ? Transmission Efficiency and
Throughput ?
2
Causes of errors

• Errors are caused by
• various kinds of surrounding noise which disturbs
  the signal going through a medium like copper,
  coaxial cable, etc.
• properties of the medium
• attenuation distortion (high frequencies lose
  power more rapidly than low frequencies)
• delay distortion (different frequencies travel
  through the medium at different speeds)

3
Error Prevention

• Shielding
• Relocating cables
• Conditioning (carriers guarantee the maximum
  number of errors that can occur)
• C-type conditioning compensates for attenuation
  and delay distortions.
• D-type conditioning improves signal to noise
  ratio.

4
ERROR DETECTION AND CONTROL PARITY CHECKING
Single Parity bit Total number of 1 bits must
always be even. V 0110101 (7-bit ASCII code).
Since, the number of 1s is even, add a 0 as the
eighth bit. Therefore, 8-bit representation of V
is 01101010. Similarly, W 0001101 (7-bit ASCII
code) Since, the number of 1s is odd, add a 1 as
the eighth bit. Therefore, 8-bit representation
of W is 00011011. (In odd parity system, total
number of 1 bits is always odd.) What is the
drawback with the single parity method?
5
CYCLICAL PARITY CHECKING
This method requires two parity bits per
character. Assuming six bits of code (bits 1
through 6) add two parity bits (bits 7 and 8)
such that bit 7 is the parity for bits 1, 3 and
5, while bit 8 is the parity for bits 2, 4, and
6. Again, total number of ones is even in both
cases.
0 1 1 0 0 1 1 0
Parity 2
Parity 1
How is this method better than having a single
parity bit?
6
M-of-N Codes The code is designed in such a way
that there will always be M 1s and N-M 0s in
each valid character of the code. Example 4-of-8
Code (from IBM) In this 8-bit code there must be
exactly 4 ones and 4 zeros. Valid
characters Invalid characters 00001111 0000
0111 01011010 11100000 00011110
7
Hamming codes
The table above shows how the character 1100110
is converted into its hamming code equivalent.
Even parity is used in this case.
Bit 1 checks 1, 3, 5, 7, 9, 11 Bit 2 checks
2, 3, 6, 7, 10, 11 Bit 4 checks 4, 5, 6, 7 Bit
8 checks 8, 9, 10, 11
Hamming codes are Forward error correcting codes
8

• Cyclical Redundancy Check (CRC)
• This is applied to an entire block of data in
  synchronous
• communication.
• A 16-bit (or more commonly 32-bit) number is
  calculated from the entire block, and
• attached to the end of the block by the sender.
• The receiver performs a similar calculation and
  compares the
• 16-bit value to see if it is the same. If they
  are not the same,
• it indicates an error in the transmission.
• This is a highly reliable scheme with almost 100
• error detection capability.

9
Transmission Efficiency and throughput

• Transmission efficiency is defined as
• In asynchronous transmission, efficiency 70
• In synchronous transmission, efficiency is much
  higher

10
Throughput
Throughput number of information bits received
per second after also accounting for
retransmissions due to errors.
Efficiency 80 Error rate 1 Modem speed
9600 bits per second Throughput 9600 x 0.80 x
(1 0.01) 7603.2 bits per
second This is also called transmission rate of
information bits. Note Throughput is less than
efficiency.
11
Network Architecture
Defines 1) The way communications functions are
divided into layers. 2) Protocols, standards
and messages at each layer. Objective of the
layered approach 1) Each layer performs one set
of functions. 2) Each layer isolates the layers
above it from the complexities below Protocols in
each layer are the set of rules agreed to and
followed by both parties for successful
communication.
12
Layered Network Architecture

• Several layers are involved in data
  communications (7 in OSI, 4 in TCP/IP)
• The most important layers (and the ones in
  TCP/IP) are
• Application layer handles the details of
  particular applications (e.g., Telnet, Ftp, SMTP,
  SNMP).
• Transport layerprovides reliable flow of data
  between end system hosts for the application
  layer.
• Network layer performs addressing and routing.
• Link Layer responsible for error control, flow
  control, message delineation, link management
  (media access control). Also called network
  interface layer.
• General principle division of work across layers.

13
OSI Layered Protocol
Host Computer
Host Computer
Application layer
Application layer
Presentation layer
Presentation layer
Session layer
Session layer
Transport layer
Transport layer
Network layer
Network layer
Network layer
Network layer
Data Link layer
Data Link layer
Data Link layer
Data Link layer
Physical layer
Physical layer
Physical layer
Physical layer
(Intermediate node)
14
What if your Web-browser used an Implementation
of the OSI model?
15
(No Transcript)
16
SERVER
Application Programming Interface
Application layer
Presentation layer
Session layer
Transport layer
Network layer
Data Link layer
Physical layer
17
OSI Layered Protocol
Host Computer
Host Computer
Application layer
Application layer
Presentation layer
Presentation layer
Session layer
Session layer
Transport layer
Transport layer
Network layer
Network layer
Network layer
Network layer
Data Link layer
Data Link layer
Data Link layer
Data Link layer
Physical layer
Physical layer
Physical layer
Physical layer
(Intermediate node)
18
The Level 3 Approach
Network Layer
Primary Attribute
3
Addressing
QoS
2
Multiplexing
Low Error Rate
Fault Tolerance
High Capacity
1
Physical Medium
19
Architecture Example

• The actual flow of data is between WWW
  applications.
• The apparent flow of data is peer to peer across
  the network.
• The operation of the application processes is
  independent of
• the underlying communications and network
  technologies
• -- hence a communications architecture.

20
Layers in protocols
Source Kurose and Ross (2001), Computer
Networking A Top-Down Approach Featuring the
Internet
21
Implementing an Architecture

• Each layer appends its own header to the
  application data.
• At the receiving end, each layer strips off the
  corresponding header.

22
Data Link protocols
Need a reliable way of exchanging information at
data link layer

• BSC (Binary Synchronous Communications)
• SDLC (Synchronous Data Link Control)
• HDLC (High Level Data Link Control)
• Protocol Features and Issues
• Communications line control (polling/selecting)
• Framing
• Addressing
• Synchronization
• Data transparency
• Error control
• Flow control
• Fragmentation and reassembly

23
Multipoint SDLC network
24
SDLC Frame
Frame Check Sequence
Information
Ending Flag
Begin Flag Address Control

8 bits
01111110
8 or 16 bits
Variable length
16 bits
01111110
25
How Data Link Protocol Works(Automatic Repeat
Request - ARQ - method)

• Stop and wait ARQ
• Sender stops and waits for response from receiver
  after each packet
• Receiver sends ACK if no errors in message
• Receiver sends NACK if errors in message.
• This is a half-duplex method used in BSC protocol.

26
Stop and wait ARQ
B
A
frame 0
Ack1
frame 1
Ack0
In this case, there are only two frames numbered
0 and 1
27
Protocol for error correction(Automatic Repeat
Request)

• Continuous ARQ
• sender does not wait for response from receiver
  after each packet
• receiver asks for retransmission of erroneous
  packets.
• This is a full-duplex method.
• It is also called sliding window protocol.
• It is used in SDLC protocol.

28
B
A
Continuous ARQ
Frame 0 received okay
Frame 1 received okay
Frame 2 received okay
Frame 3 received okay
Frame 4 received okay
Frame 5 not received
Frame 6 received okay
Frame 7 received okay
Frame 5 received okay
Frame 6 received okay
Frame 7 received okay
Note Received Receipts will go back from B to A
(not animated)