Error Detection, Correction and Control

1
Error Detection, Correction and Control

• Rong Wang
• CGS3285
• Spring 2004

2
RECOMMENDED READING

• From textbooks
• Page 117 123 of Data Communications From
  Basics to Broadband, 3rd Edition by William J.
  Beyda (ISBN 0-13-096139-6)
• Chapter 10 11 of Data Communications and
  Networking, 3rd Edition, Behrouz A. Forouzan
  (ISBN 0-07-251584-8)

3
TRANSMISSION ERROR

• A bit is altered between transmission and
  reception
• Where does errors come from?
• Electrical Interference
• Power surges
• Noise, static, echoes, cross-talk
• Signal degradation
• Type of error
• Single-bit error isolated, does not affect
  nearby error (e.g., white noise)
• Burst error a cluster of bits involved (faded
  signal)

4
ERROR DETECTION METHODS

• Echo Checking
• Receiver returns copy of data back to sender
• Ignore Parity Checking
• Used if sender or receiver unable to process
  parity bits
• Primitive Parity Checking
• Mark - Always send a 1
• Space - Always send a 0
• 12.5 chance of detecting a 1-bit error
• Simple Parity Checking
• Even - Make of 1s in character even
• Odd - Make of 1s in character odd
• 50 chance of detecting a 1-bit error

5
ASCII CHARACTER TRANSMISSION WITH DIFFERENT
PARITIES
6
EVEN PARITY CHECKING AND BIT ERRORS
7
ERROR DETECTION METHODS (cont.)

• Multi-bit Parity
• Cyclical Parity
• Uses two (2) parity bits
• Fooled only by errors in certain bit combinations
• Increases overhead in both data transmitted and
  processing time

8
FIGURE 6-3 CYCLICAL ODD PARITY CHECKING
9
ERROR DETECTION METHODS (cont.)

• Multi-bit Parity (continued)
• Hamming Code (forward error correction)
• Form of forward error correction
• Can detect and repair single-bit errors
• Uses seven (7) data bits
• Bit Positions 3, 5, 6, 7, 9, 10, 11
• Uses four (4) parity bits
• Bit Position 1 Bits 3, 5, 7, 9, 11
• Bit Position 2 Bits 3, 6, 7, 10, 11
• Bit Position 4 Bits 5, 6, 7
• Bit Position 8 Bits 9, 10, 11
• Sum of parity bits with errors is bad data bit
• Increases overhead in both data transmitted and
  processing time

10
FIGURE 6-4 HAMMING w/ EVEN PARITY
11
ERROR DETECTION METHODS (cont.)

• Synchronous Block Methods
• Checksum
• Use simple arithmetic to generate a checksum
  remainder which is appended to end of message
• Sender creates checksum as block is transmitted
  then appends to end of block
• Receiver generates checksum as block is received
  then compares with senders checksum
• If error, receiver asks for re-transmission of
  block
• Requires checksum and block sequence number
• Cyclical Redundancy Check (CRC)
• Similar to check sum but uses more complex
  mathematical formulas to catch/exaggerate errors

12
ERROR DETECTION / CORRECTION SUMMARY

• Error detection correction designed to preserve
  the INTEGRITY of transmitted data -- not SECURITY
• Detection is relatively easy
• Correction is harder
• Retransmission is simplest approach
• Both sender and receiver must use same
  detection/correction method
• Type of method and algorithms specified by
  protocol

13
DATA LINK ERROR CONTROL (CONT)

• Error control
• Based on automatic repeat request, which is the
  retransmission of data
• Type of error handled
• Lost frame
• Damaged frame
• Techniques for error control
• Error detection
• Positive acknowledgement
• Retransmission after timeout
• Negative acknowledgement and retransmission
• Note data link layer divides the stream of bits
  received into manageable data units called
  frames.

14
DATA LINK ERROR CONTROL (CONT)

• Stop Wait ARQ (Half Duplex)
• Receiver responds to each block (ACK / NAK /
  WACK)
• Sender waits for response before sending further
  blocks
• Continuous ARQ (Full Duplex)
• Blocks sent continuously without waiting for
  confirmation
• If error message returned to sender,
• Types of continuous ARQ
• Go-Back-N restart transmission at the Nth frame
• Selective Repeat retransmit the selected frame
  only

15
STOP-AND-WAIT ARQ

• Note To avoid duplication, data frames are
  alternately labeled with 0 or 1, thus
  acknowledgement (ACK) frames are also numbered 0
  and 1 alternately.
• Normal Operation
• Sender sends frame 0 and waits to receive ACK 1
• When ACK 1 is received, it sends frame 1 and then
  waits to receive ACK0
• repeat from step 1 and 2 and so on.
• The ACK must be received before the time set for
  each frame expires

16
STOP-AND-WAIT ARQ (CONT)

• Lost or Damaged Frame
• A lost or damaged frame is handled in the same
  way by the receiver
• When the receiver receives a damaged frame, it
  discards it
• The receiver remains silent about a lost frame
  and keeps its value of R
• After the timer at the sender site expires,
  another copy of frame 1 is sent

17
STOP-AND-WAIT ARQ (CONT)

• Lost or Damaged Acknowledgement
• If the sender receives a damaged ACK, it discards
  it.
• When the timer for a frame expires, the sender
  retransmits the frame.

18
STOP-AND-WAIT ARQ (CONT)

• Delayed Acknowledgement
• An acknowledge can be delayed at the receiver or
  by some problem with the link
• If receiver receives a duplicated frame, it
  discards the duplicated frame

19
STOP-AND-WAIT ARQ (CONT)

• Piggybacking
• A method to combine a data frame with an
  acknowledgement in case both sender and the
  receiver have data to send.

20
GO-BACK-N ARQ

• Sequence Numbers
• Frame from a sending station are numbered
  sequentially. If the header of the frame allows m
  bits fro the sequence number, the sequence
  numbers range from 0
• Sender Sliding Window
• Sender uses window to hold the outstanding frames
  until they are acknowledged
• Frames to the left of the window are those that
  have already been acknowledged and can be purged
• Frames to the right of the window can not be sent
  until the window slides them

21
GO-BACK-N ARQ (CONT)

• Receiver Sliding Window
• Size of the window is always 1
• The receiver is always looking for a specific
  frame to arrive in a specific order

22
GO-BACK-N ARQ (CONT)

• Control Variables
• The sender has three variables
• S holds the sequence number of the recently sent
  frame
• holds the sequence number of the last
  frame in the window
• Holds the sequence number of the first
  frame in the window
• Size of the window
• The receiver only has one variable R, that holds
  the sequence number of the frame it expects to
  receive

23
GO-BACK-N ARQ (CONT)

• Normal Operation

24
GO-BACK-N ARQ (CONT)

• Damaged or Lost Frame

25
GO-BACK-N ARQ(CONT)

• Sender window size

26
GO-BACK-N ARQ(CONT)

• Problem
• No out-of-order frames
• Not efficient
• Multiple frame may be retransmitted

27
SELECTIVE REPEAT ARQ

• Only damaged frame is resent
• Sender and receiver windows
• Size of the senders window must be one half of
• Size of the receivers window must be same as the
  senders
• Negative acknowledgement (NACK) is used to define
  damaged frame

28
SELECTIVE REPEAT ARQ (CONT)

• Lost Frame

29
SELECTIVE REPEAT ARQ (CONT)

• Sender window size