ERROR CONTROL; DETECTION & CORRECTION - If even numbers of bits are received in error, the

parity checker will not detect it because when the

Data/Digital Communication Error can be generally logic condition of an even number bits is
classified into: changed, the parity remains the same
1. Single Bit Error– only one bit within the given data string - Odd parity is usually use, does not identify which
is in error character is in error
- Affects only one character within the message
2. Multiple Bit Error – when two or more nonconsecutive
bits within a given data strings are in error Ex. ASCII Letter C is 43 hex or P1000011 binary:
- Affects more character in a message
3. Burst Error – when two or more consecutive bits within a P is the Parity: for ODD Parity (P=logic 0)
given data string are in error For EVEN Parity (P=logic 1)

Two Basic Types of Error: - Other forms of parity

1. Lost Message – one that never arrives at the destination a. Space parity – parity bit always a 0
- arrives but has been damaged and unrecognizable b. Marking parity – parity bit always a 1
2. Damaged Message – arrives at the destination but has one or c. No Parity – the parity bit is not seen or checked
two errors. d. Ignored Parity – parity bit is always zero and is
ignored
Probability of Error - Allows receiver that are incapable
-the theoretical (mathematical) expectation of the rate at which of checking parity to communicate with the devices
error will occur that uses parity

Bit Error Rate a.2.Checksum

- The actual historical record of a systems performance - The data within a message is summed together to
produce an error checking character (checksum)that is
Error Control can be divided into: appended at the end of the message
- Receiver replicates the summing operation and
I. ERROR DETECTION determines its own sum and checksum character for
- Process of monitoring data and determining when the message.
transmission errors have occurred - It is then compared to the checksum appended to the
- Indicated when an error has occurred to prevent message; no errors are transmitted if they are the
undetected errors from occurring same, errors occurred if the two values differ.

Most Common Error Detector - Five Primary Ways of Calculating a Checksum

a.2.1.Check Character Checksum
A. REDUNDANCY - A decimal value is assigned to each character
- Each data is sent multiple times which are then and they are added together to produce the
compared at the receivers end checksum character appended at the end of the
- Character redundancy – for data with one character message
- Message redundancy – if the data is the entire - Ex. Code 39 and POSTNET bar code
message
- An error will occur if the data is not received twice a.2.2.Single Precision Checksum
in succession in exactly the same order - Checksum is calculated by simply performing
binary addition of the data within the message
Redundancy Checking - If the sum of the data exceeds 2n-1 where n is
- Adding bits for the sole purpose of detecting the number of bits in each character, a carry-
errors out occurs and is ignored

4 Basic Types of Redundancy Checking Example:

a.2.2.Determine the single-precision checksum for the
a.1. Vertical Redundancy Checking following five-character ASCII message: HELLO
- Used for asynchronous data communication
systems
- Referred to as character parity or parity
- A single bit is added (parity bit) is added to each a.2.3.Double Precision Checksum
character to force the total number of logic 1s in - computed in the same manner as with single-
the character, including the parity bit to either an precision except the checksum is 2n bits long
odd number (odd parity) or an even number (even - ex. If the data is comprised of 8-bit character, the
parity) check sum would be 16 bits
- Primary advantage is its simplicity - reduces the probability of producing erroneous
checksum
 - Mathematically, CRC is expressed as:
a.2.4.Honeywell Checksum
- 2n bits long where the checksum is based on the
interleaving consecutive data words to form 𝐺(𝑥)
= 𝐺 𝑥 + 𝑅(𝑥)
double-length words are summed together to 𝑃(𝑥)
produce a double-precision checksum
Where: G(x) = message polynomial
Example a.2.4: P(x) = generator polynomial
Determine the Honeywell checksum for the R(x) = remainder
following four-character ASCII message: HELP
How do we compute R (the check bits)?
1. Choose a generator string G of length r+1 bits
2. Choose R such that T is a multiple of
a.2.5. Residue Checksum G (T = A*G, for some A)
- virtually identical to the single-precision 3. Now when T is divided by G there will be no
checksum except for the way carry bit are remainder =>no error
handled 4. Let R = remainder of M 2r/G and T will be a multiple
-the carry bit is wrapped around and added to the of G
LSB of the sum, adding complexity 5. Choice of G is a critical parameter for the performance
of a CRC
6. Let T’ be the received sequence
a.3. Longitudinal Redundancy Checking 7. Divide T’ by G
- uses parity to determine if a transmission error has – If remainder = 0 assume no errors
occurred within a message – If remainder is non zero errors must have occurred
- sometimes called message parity
- result of XORingthe “characters” that comprise the Example A.4:
message
- even parity is generally used
- sometimes called Horizontal Redundancy Checking B. ECHOPLEX / ECHO CHECKING
(HRC) - Receiving devices retransmit received data back to the
- bit sequence is often called Block Check Sequence transmitting device
(BCS) or Frame Check Sequence (FCS) - Requires full-duplex transmission
- detects 95% to 98% error - a wrong character will be seen in the transmitting
- does not identify which character is in error station if an error will occur then the operator sends
back-space and remove the erroneous character and
Example a.3: then type and resend the correct character
Determine the VRC and the LRC for the following ASCII-
encoded message CAT. Use odd parity for VRCs and even C. EXACT-COUNTING ENCODING
parity for the LRC - The number of 0’s and 1’s in each character is the
same
- Ex. Code 39 Bar Code – 9-bit w/ 3 logic1bits
a.4. Cyclic Redundancy Checking - POSTNET Bar Code – 5-bit w/ 2 logic1 bits
- most reliable technique for error detection -ARQ character – 7-bit w/ 3 logic1 and 4 logic 0
- a convolutional coding - Error can be recognized if a certain number of 1 or
- CRC-ITU : European Standard 16-bit CRC-generating zero in a character is lacking
polynomial
- CRC-32 : for point-to-point synchronous data
transmission
- CRC-12 : 12-bit redundancy code used for
transmission of data stream comprised of six-byte
- CRC-16 : 16 bits are used for block check sequence
- The entire data stream is treated as long continuous
binary number
- Systematic code
- Written as (n, k) ; n=bit length of transmission, k=bit
length of the message
- Detects all single bits error and all double-bit error
provided that the divisor contains at least three logic
1, all odd number of bit error provided that the bit
divisor is 11, all error burst of 16 or less, and 99.9%
of error burst greater than 16 bits long
II. ERROR CORRECTION Sample Problem:

A. RETRANSMISSION(BACKWARD ERROR For a 12-bit data string 101100010010, determine the number
CORRECTION) of Hamming bits required, arbitrarily place the Hamming bits
- when receive station request the transmit station to into the data string, determine the logic condition of each
resend a message (portion of the message) when the hamming bit, assume an arbitrary single-bit transmission error,
message is received in error that is often called ARQ ( and prove that the Hamming code will successfully detect an
automatic retransmission request) error.

Forms of Overhead
1. acknowledgement – when the recipients of data sends
short messages back to the sender acknowledging
receipt of the last transmission,
- can either be positive
acknowledgement (successful transmission)
negative acknowledgement (unsuccessful
transmission)

2. line turnarounds–when a receive station becomes the

transmit station

Two Types of ARQ

1. Discrete ARQ –use acknowledgements to indicate the
successful and unsuccessful reception of data
- positive acknowledgement- if the message is
received correctly
- negative acknowledgement – if the message
received contains error

2. Continuous ARQ – when messages are divided into

smaller blocks or frames that are sequentially
numbered and transmitted in succession without
waiting for acknowledgement between blocks.
- selective repeat- use to call for retransmission of the
entire message or only a portion of the message

B. HAMMING CODE(FORWARD ERROR

CORRECTION)

Hamming Code –an error-correcting code used for

correcting transmission errors in synchronous data
streams
- requires the addition of overhead
message to the message increasing the length of the
transmission

Hamming Bits – must be able to identify which bit is in

error and indicate at least m + n + 1 different codes
- the number of hamming bits is indicated by

2n> m + n + 1

Where: n = number of Hamming bits

m = number of bits in each data