15
$\begingroup$

As I understand it, a hashing algorithm is some kind of encryption, just a specific kind of it. So is it wrong to say "I encrypted this value." if I "only" hash it?

I understand that it gives the wrong impression, but I think it is not absolutely wrong, or is it?

$\endgroup$
1
  • 4
    $\begingroup$ Comments that are asserting that hashing is not encryption or vice versa will be removed. Please post a well explained answer instead if your point is not covered in the existing ones. $\endgroup$
    – Maarten Bodewes
    Sep 10 '19 at 13:00
53
$\begingroup$

Encryption algorithms and hash algorithms both belong to the realm of cryptography but are two different things: Encryption doesn't contain hash functions.

As stated on Wikipedia:

In cryptography, encryption is the process of encoding a message or information in such a way that only authorized parties can access it and those who are not authorized cannot.

An encryption algorithm turns some data into a ciphertext which can later be decrypted again to its original content by using the encryption key.

If you generate a hash of some data it can't be "reversed" back into its original content, because a hash function is (theoretically) a one-way function with loss of data.

$\endgroup$
3
  • $\begingroup$ Comments are not for extended discussion; this conversation - which is mainly about the word "theoretically" in the answer - has been moved to chat. $\endgroup$
    – Maarten Bodewes
    Sep 11 '19 at 10:48
  • 2
    $\begingroup$ A small note on terminology: historically cryptography was only about encryption / decryption. Cryptology was a more broad term, that also included analysis. Nowadays they are synonymous, and both are broadened to include secure hashing, random number generation, signature creation/verification etc. etc. etc. and the analysis thereof. Hence our site name :) $\endgroup$
    – Maarten Bodewes
    Sep 11 '19 at 13:08
  • $\begingroup$ The last part refers to cryptographic hash functions. In general they don't need to be one-way and could allow for (partly) reconstruction. $\endgroup$
    – jjj
    Oct 13 at 14:04
20
$\begingroup$

Encryption implies that with the appropriate key, it is possible to decrypt and recover the original message. Which (in general) is not possible from a hash.

Thus “I will encrypt” is not adequate if one is going to hash.

While it is possible to construct hashes from encryption primitives (such as block ciphers), and vice versa, they are different beasts.

$\endgroup$
2
  • $\begingroup$ Why did you say 'in general'? $\endgroup$ Sep 11 '19 at 10:29
  • 1
    $\begingroup$ @Begueradj Non cryptographic hashes may not be one way. Trying all possible inputs could lead you to find the correct input. $\endgroup$
    – Maarten Bodewes
    Sep 11 '19 at 12:20
13
$\begingroup$

I understand that it gives the wrong impression, but I think it is not absolutely wrong, or is it?

It is actually. A hash algorithm computes a 'fingerprint' if you will of the input. So just as a fingerprint identifies you, a hash identifies the input document.

But just as an entire human being cannot be recreated from just a fingerprint, so the original document cannot be (easily) recreated from the hash. This is the fundamental distinction with an encryption. In encryption it's useful to decrypt, using the original encryption key. You can't with a hash, and it has no key.

$\endgroup$
3
  • 2
    $\begingroup$ so the original document cannot be (easily) recreated from the hash – For most messages, it's fundamentally impossible. $\endgroup$
    – forest
    Sep 10 '19 at 6:31
  • 2
    $\begingroup$ @forest: ...because of the pigeonhole problem. Finding a preimage is theoretically possible, but there's no way to tell which was the original preimage - multiple "original documents" can hash to the same value. See e.g. natmchugh.blogspot.com/2014/10/… - from the output hash, it is impossible to tell which input was used. $\endgroup$ Sep 10 '19 at 11:06
  • 2
    $\begingroup$ @Piskvor That's exactly the point I wanted to make. $\endgroup$
    – forest
    Sep 10 '19 at 11:08
2
$\begingroup$

I agree with other answers, but would like to add this: when talking to someone familiar with cryptography or fields using it (e.g. programming), you should indeed use the right vocabulary. But if you try to summarize what you are doing to a client, I would use neither and say something like "we store our password in a secure way" and eventually provide details somewhere else.

$\endgroup$
4
  • 1
    $\begingroup$ That might give a wrong impression of what is actually done. There is a substantial difference between storing a password and not storing a password (just the hash), which must be understood even by non-technical people. Any system that stores passwords brings potential security risk, and if you are a customer looking for developers, you must know the difference (or you will be left with a risky system made by developers who know nothing about hashing). $\endgroup$
    – IS4
    Sep 10 '19 at 11:17
  • 1
    $\begingroup$ Agreed, it may be not very well phrased, but the point would be: know what you do, and make it clear for others without the need to be too specific. If the explanation is for a final customer, I don't believe he needs to know what a hash is. $\endgroup$
    – Lou_is
    Sep 10 '19 at 11:24
  • $\begingroup$ I would argue that this is not an answer within the scope of crypto.se. We do not really deal with customer relations here. $\endgroup$
    – Maeher
    Sep 10 '19 at 18:07
  • 3
    $\begingroup$ @Lou_is: I think that most customers are OK with a little bit of jargon, as long as the meaning is clear enough. Something like "We only store secure hashes of passwords, so even if someone managed to steal the entire database, they still wouldn't have the passwords" should be fine. $\endgroup$
    – ruakh
    Sep 11 '19 at 0:18
1
$\begingroup$

I'm assuming this is something the user will get to see later? If that's the case I'd recommend just writing "hash" with a short explanation of what that means.

Something like:

Hash means that whatever it's used for is uniquely identifiable, but not reconstructible.

Even though in rare cases hashes overlap and therefore aren't 100% unique, if you're target audience isn't tech savvy - I'd just leave that out because it would add unnecessary complexity to the statement.

$\endgroup$
0
1
$\begingroup$

You will hurt yourself if you claim it is encryption, because (especially for password hashing) if a review is done password encryption (I.e. reversible encryption) should be flagged as a policy violation. So it is better to never use a incorrect term (in this assumed context). Otherwise you will have to do some back and forth to explain the actual method used.

Of course if you are not in a enterprise setting this might not apply to you (but there is still nothing gained from using the wrong terms)

$\endgroup$
0
0
$\begingroup$

I just wanted to add a confusing, and weird phenomena, where laypeople,(e.g. journalists), very often conflate encryption and hashing.

It often happens when data leaks are reported and the news article says something like "encrypted passwords were leaked". What they mean 99% of the time is that "hashed passwords were leaked".

Here's a likely example from today's Twitch data leak:

Some Twitter users have started making their way through the 125GB of information that has leaked, with one claiming that the torrent also includes encrypted passwords, and recommending that users enable two-factor authentication to be safe.

$\endgroup$
6
  • $\begingroup$ Well, some really use encryption for passwords, not common though. Most of the time it should be as hashed as you said. Also some use irreversible encryption for password hashing, too. You might consider to add this, too. $\endgroup$
    – kelalaka
    Oct 6 at 17:08
  • $\begingroup$ @kelalaka Do you mean some sites encrypt rather than hash? (because, yikes!). Or do you mean they encrypt the already hashed password? If they encrypt the hash, I guess that’s possible but I also presume a hacker who get into the db would likely be able to get to a single environment variable to decrypt the hashes, although it would be an extra inconvenience I guess, therefore offering a little extra protection. $\endgroup$
    – stevec
    Oct 6 at 17:12
  • $\begingroup$ Yes, encryption, they can use HSM to store the key, though still, the attacker can use them, at least in the attack time. Remember, Facebook stored them openly! You can find strange usages in the wild. $\endgroup$
    – kelalaka
    Oct 6 at 17:19
  • $\begingroup$ @kelalaka please feel free to edit the answer to include encryption. I wasn’t even aware sites used an alternative method other than plain hashing, I’m very interested to learn more. And I didn’t know about the Facebook situation, will google/read now. $\endgroup$
    – stevec
    Oct 6 at 17:29
  • $\begingroup$ mentalfloss.com/article/577690/… $\endgroup$
    – kelalaka
    Oct 6 at 18:10
-1
$\begingroup$

Typically we require encryption be reversible, so it can be used as a communication system.

Mathematically, when we talk about a cryptographic system, we are focusing on invertible functions. This change in viewpoint is important, because it opens up the possibilities for how we can encrypt information.

For example, we can use a linear function (affine transformation) as an encryption system:

https://youtu.be/1S92scw5zIg?list=PLKXdxQAT3tCssgaWOy5vKXAR4WTPpRVYK

$\endgroup$
2
  • 1
    $\begingroup$ Welcome to crypto.stackexchange - Can you please summarize the content of the link, or mention what the content is/why it's relevant? If the link ever dies or changes, it greatly diminishes the utility of the answer. $\endgroup$
    – Ella Rose
    Sep 9 '19 at 17:01
  • 3
    $\begingroup$ There's not a single word "hash" in this answer. $\endgroup$ Sep 10 '19 at 1:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.