By Iliya Garakh, CTO in Cryptography — Nov 5, 2021

What is password hashing and salting?

Cryptography is both beautiful and terrifying. Perhaps a bit like your ex-wife. Despite this, it represents a vital component of day-to-day internet security; without it, our secrets kept in the digital world would be exposed to everyone, even your employer. I doubt you’d want information regarding your sexual preferences to be displayed to the regional sales manager while at an interview with Goldman Sachs, right?

Computers are designed to do exactly what we ask them to do. But sometimes there are certain things that we don’t want them to do, like expose your data through some kind of backdoor. This is where cryptography comes into play. It transforms useful data into something that can’t be understood without the proper credentials.

Let’s take a look at an example. Most internet services need to store their users’ password data on their own servers. But they can’t store the exact values that people input on their devices because, in the event of a data breach, malevolent intruders would effectively gain access to a simple spreadsheet of all usernames and passwords.

This is where ‘Hash’ and ‘Salt’ help us a lot. Throughout this article, we’re going to explain these two important encryption concepts through simple functions in Node.JS.

What is a ‘hash’?

A ‘hash’ literally means something that has been chopped and mixed, and originally was used to describe a kind of food. Now, chopping and mixing are exactly what the hash function does! You start with some data, you pass it through a hash function where it gets whisked and chopped, and then you watch it get transformed into a fixed-length value (which at first sight seems pretty meaningless). The important nuance here is that, contrary to cooking, an input always produces a corresponding output. For the purposes of cryptography, such a hash function should be easily computable and all values should be unique. It should work in a similar way to mashing potatoes – mashing is a one-way process; the raw potato may not be restored once it has been mashed. Indeed, the result of a hash function should be impenetrable to computer-led reverse engineering efforts.

These properties come in handy when you’re looking to store user passwords on a database – you don’t want anyone to know their real values.

Let’s implement a hash in Node.JS!

First, let’s import the createHash function from the built-in ‘crypto’ module:

const { createHash } = require ('crypto');

Next, we ought to define the module that we’re naming as the ‘hash’ (which takes a string as the input, and returns a hash as the output):

function hash(input) {
    return createHash();
}

We also need to specify the hashing algorithm that we want to use. In our case, it will be SHA256. SHA stands for Secure Hash Algorithm and it returns a 256-bit digest (output). It is important to architect your code so it is easy to switch between algorithms because at some point in time they won’t be secure anymore. Remember, cryptography is always evolving.

function hash(input) {
    return createHash('sha256');
}

Once we call our hashing function, we may call ‘update’ with the input value and return the output by calling ‘digest’. We should also specify the format of the output (e.g. hex). In our case, we’ll go with Base64.

function hash(input) {
    return createHash('sha256').update(input).digest('base64');
}

Now that we have our hash function, we can provide some input, and console log the result.

let youShallNotPassPass = 'admin1234';
const hashRes1 = hash(youShallNotPassPass);
console.log(hashRes1)

Here’s our baby hash:
rJaJ4ickJwheNbnT4+I+2IyzQ0gotDuG/AWWytTG4nA=

So, how can we use this long, convoluted string of numbers, letters, and symbols? Well, now it’s easy to compare two values while operating with only hashes.

let youShallNotPassPass = 'admin1234';
const hashRes1 = hash(youShallNotPassPass);
const hashRes2 = hash(youShallNotPassPass);
const isThereMatch = hashRes1 === hashRes2;
console.log(isThereMatch ? 'hashes match' : 'hashes do not match’)

As long as hash values are unique object representations, they can be useful for object identification. For example, they might be used to iterate through objects in an array or find a specific one in the database.

But we have a problem. Hash functions are very predictable. On top of that, people don’t use strong passwords that often, so the hacker may just compare the hashes on a database with a precomputed spreadsheet of the most common passwords. If the values match – the password is compromised.

Because of this, it’s insufficient to just use a hash function to store unique ids on a password database.

And that’s where our second topic makes an entrance – Salt.

‘Salt’ is a bit like the mineral salt that you would add to a batch of mashed potatoes – the taste will definitely depend on the amount and type of salt used. This is exactly what salt in cryptography is – random data that is used as an additional input to a hash function. Its use makes it much harder to guess what exact data stands behind a certain hash.

So, let’s salt our hash function!

First, we ought to import ‘Scrypt' and ‘RandomBytes’ from the ‘crypto’ module:

const { scryptSync, randomBytes } = require('crypto');

Next, let’s implement signup and login functions that take ‘nickname’ and ‘password’ as their inputs:

function signup(nickname, password) { }
function login(nickname, password) { }

When the user signs up, we will generate a salt, which is a random Base64 string:

const salt = randomBytes(16).toString('base64');

And now, we hash the password with a 'pinch' of salt and a key length, which is usually 64:

const hashedPassword = scryptSync(password, salt, 64).toString('base64');

We use ‘Scrypt’ because it’s designed to be expensive computationally and memory-wise in order to make brute-force attacks unrewarding. It’s also used as proof of work in cryptocurrency mining.

Now that we have hashed the password, we need to store the accompanying salt in our database. We can do this by appending it to the hashed password with a semicolon as a separator:

const user = { nickname, password: salt + ':' + hashedPassword}

Here’s our final signup function:

function signup(nickname, password) {
    const salt = randomBytes(16).toString('base64');
    const hashedPassword = scryptSync(password, salt, 64).toString('base64');
    const user = { nickname, password: salt + ':' + hashedPassword};
    users.push(user);
    return user;
}

Now let’s create our login function. When the user wants to log in, we can grab the salt from our database to recreate the original hash:

const user = users.find(v => v.nickname === nickname);
const [salt, key] = user.password.split(':');
const hash = scryptSync(password, salt, 64);

After that, we simply check whether the result matches the hash in our database. If it does, the login is successful:

const match = hash === key;
return match;

Here is the complete login function:

function login(nickname, password) {
    const user = users.find(v => v.nickname === nickname);
    const [salt, key] = user.password.split(':');
    const hash = scryptSync(password, salt, 64).toString('base64');
    const match = hash === key;
    return match;
}

Let’s do some testing:

//We register the user:
const user = signup('Amy', '1234');

//We try to login with the wrong pass:
let isSuccess = login('Amy', '12345');
console.log(isSuccess ? 'Login success' : 'Wrong password!')

//Wrong password!
//We try to login with the correct pass:
isSuccess = login('Amy', '1234')
console.log(isSuccess ? 'Login success' : 'Wrong password!')

//Login success

Our example, hopefully, has provided you with a very simplified explanation of the signup and login process. It’s important to note that our code is not protected against timing attacks and it doesn’t use PKI infrastructure to check hashes, so there are plenty of vulnerabilities for hackers to exploit.

Cryptography itself can be described as the constant war between hackers and cryptographic engineers. Or, that familiar legal battle with your ex-wife over her maintenance payments. After all, what works today may not work tomorrow. A proof of MD5 hash algorithm vulnerability is a very good example.

So if your task is to ensure your users’ data privacy, be ready to constantly update your functions to counteract the recent ‘breakthroughs’.