Hashes: collisions vs deterministic

bermovick

I just cannot understand how the definitions of these 2 things are not actually explaining the same thing.

From the official cert guide, a collision is when 2 distinct inputs entered into a hash function produce identical outputs. IE a collision is when hash(A) = C and hash(B) = C if A != B.

Being deterministic is when a hash function generates 2 hashes that are different, the 2 inputs were different. IE non-deterministic would mean the hash generates the same hash with 2 different inputs: hash(A) = C and hash(B) = C and A != B.

Is my logic wrong, or are the 2 'characteristics' actually just the same thing worded differently? I'd write it off, but this information is referenced several times so it seems to be somewhat important; pages 467, 469, 475, and 477.

Chris:/*

Deterministic means that a value must always give the same hash value period.

Collision means that different values can end up equaling the same hash despite being different.

bermovick

so it's sortof the opposite.

the first; if hash(A) = C and hash(B) = D, then if A and B are different, C and D should be different.

The second: if hash(A) = C and hash(B) = D, then if C and D are different, A and B should be different.

That makes quite a bit more sense in practice (this book seems to make things more difficult/confusing than they should be / need to be).

Chris:/*

Are you reading the official CCNA Security Study Guide?

Grab the a good Security+ book (the Security+ forum has a good one currently being spoken of). Learn the concepts from that book first and then study the Cisco stuff for the exam. Cisco often does not clearly explain what they are trying to get across.

medic

Another option is Google. Just google a concept you are having difficulty with and you will usually find a good explaination of it somewhere.

medic

Also be aware the Official Exam guide is full of typos. If something doesn't make sense, it is usually because it is a mistake in the book.