Error Detection & Handling in Communication Protocols

When data travels between devices, corruption can happen due to noise, interference, timing errors, or synchronization issues.
That’s why error detection and handling is a core feature of almost ALL communication protocols.

Table of Contents

This part explains how protocols detect, report, and recover from errors, using clear examples and ASCII diagrams.

1. Why Errors Happen in Communication

Common causes:

Electrical noise in wires
Crosstalk between signals
Clock mismatch
Packet collision
Signal attenuation (long distance)
EMI (Electromagnetic Interference)
Buffer overflow in receivers

2. Types of Errors

Error Type	Description	Example
Single-bit error	Only 1 bit flips	1010 → 1000
Burst error	Multiple consecutive bits corrupted	1010110 → 1110100
Packet loss	Entire packet missing	No data received
Out-of-order	Packets received in wrong order	1,3,2,4
Duplicate packets	Same data repeated	2,2,3,4

3. Common Error Detection Techniques

3.1 Parity Bit (Simple, used in UART, RAM)

Adds 1 extra bit to make total number of 1s either EVEN or ODD.

Example (Even parity):

Data:  1010  -> 2 ones (even) -> Parity = 0<br>Sent:  10100<br>

Data:  1010  -> 2 ones (even) -> Parity = 0<br>Sent:  10100<br>

Pros: Simple
Cons: Detects only single-bit errors

3.2 Checksum (Used in IP, TCP, UDP)

Adds all bytes and keeps the sum.

Example:

  0x10
+ 0x22
------
  0x32  -> checksum sent

Receiver:

Calculates again
Compares with checksum

Pros: Can detect some burst errors
Cons: Not very strong compared to CRC

3.3 CRC (Cyclic Redundancy Check) – Used in Ethernet, USB, PCIe

Most powerful and widely used.

Treats data as a polynomial, divides by a generator polynomial, and sends the remainder.

ASCII Representation:

[Data Bits] / [Generator Polynomial] = Remainder (CRC)

Pros: Detects burst errors up to 32 bits (CRC-32)
Cons: Requires more hardware/software logic

4. Error detection and Handling Approaches

Approach	What Happens
Error Detection only	Receiver discards data
Error Correction (FEC)	Receiver fixes data using math (e.g., Hamming code)
Retransmission (ARQ)	Receiver requests sender to re-send

5. Automatic Repeat Request (ARQ)

Used in TCP, Wi-Fi, LTE, etc.

5.1 Stop-and-Wait ARQ

Sender --> Frame 1 --> Receiver
Sender <-- ACK 1 ---- Receiver

Simple but slow (waits every time)

5.2 Go-Back-N ARQ

Sends multiple frames without waiting but if one fails, go back and resend all after that frame.

1 2 3 X 5
      ^
   error at 4
Resend: 4,5

5.3 Selective Repeat ARQ

Only resend the corrupted frame.
Most efficient but complex.

1 2 3 X 5
      ^
   error at 4
Resend: 4 only

6. ACK and NACK Mechanism

Term	Meaning
ACK	Acknowledged – OK
NACK	Not Acknowledged – Something went wrong

Some protocols don’t use NACK — only timeouts.

7. Timeouts and Retries

If no ACK within time, the sender retransmits.

Sender: Frame 1
(wait...)
No ACK
Resend Frame 1

8. Sequence Numbers (Important!)

Used to:

Detect lost packets
Prevent duplicates
Maintain order