HrnyGoat

ECC (Error Correcting Code) is often used in server memory, as well and network and internet connections. It is used to reduce or eliminate the chance of errors occuring in critical applications. How does it work? One method is by using parity.


Parity works by adding an extra bit to a binary number, to make all the digits add up to an even or odd number. If it adds up to an even number, it is called even parity. If the digits add up to an odd number, it is known as odd parity.


For example:

1001011
0101111

Both are 8-bit binary numbers.

Now if we want to use even parity, we add a parity bit to the end. Adding the digits together will result in an even number.

01001011 --> 4
10101111 --> 6

Or if we want to use odd parity we add a parity bit that will make the digits add up to an odd number.

11001011 --> 5
00101111 --> 5


Even parity is the most commonly used in modern applications. So now lets see how parity is used to detect errors (in this case, even parity).

Lets take our two binary numbers and change one bit.

01001011 ----> 01011011
10101111 ----> 10101101

Now lets see how they add up:

01011011 --> 5
10101101 --> 5

Notice how they now add up to odd numbers? Since we're using even parity, this would be recognized as error. It is able to detect that the number was changed during transmission.

Parity is a simple and effective way to detect errors. But it does have its shortcommings. If 2 bits are changed, it will result in an even number again and the error will not be detected. There are other, more advanced methods for ECC, such as adding multiple parity bits. But those are much more complex as well.

Crimson_Blade

nice work!

this is exactly what i needed for my college networking assignment

batista

simple google question but good thing you have it here. I was always asking myself for the difference between the nonECC and ECC but now i know. Good FaQ.

d3daiM

Isn't it true that ECC memory costs a handful?

Good FAQ, I understand it now :)