Cryptographic Hashes

A cryptographic hash function takes data and deterministically produces a fixed-length digest. The same input always produces the same digest under the same function, while even a tiny change in input is expected to produce an unrelated-looking output. Because the digest is shorter than the universe of possible inputs, collisions must exist mathematically, but a secure design makes finding useful collisions computationally infeasible.

Hashing is one-way in the security sense: the digest is not meant to be decoded back into the original message.

Preimage resistance means that given a digest, finding an input that hashes to it should be computationally hard. Second-preimage resistance means that given one specific input, finding a different input with the same digest should be hard. Collision resistance means finding any two distinct inputs with the same digest should be hard.

These are different attack goals. A hash algorithm can weaken against one property before another, which is why standards deprecate obsolete algorithms.

If block B stores the hash of block A's header, changing block A changes its hash. Block B's stored reference no longer matches, and every later descendant that depends on B is now attached to an altered chain. The hash does not physically prevent editing a file; it makes the edit detectable because commitments no longer line up.

Consensus adds the rule for deciding whether an altered history can replace the accepted one.

Older hash functions can become unsafe as attacks improve. SHA-1 is not collision-resistant enough for modern collision-sensitive uses. Blockchains may use SHA-256, Keccak-family functions, or other cryptographic hashes depending on the protocol. Security analysis must use the actual function and construction rather than treating “hashed” as a universal guarantee.

Even a strong hash cannot repair weak key management or malicious input.

Hash security becomes clearer when the three main attack goals are kept separate. In a preimage attack, an adversary starts with a digest and tries to find any input that produces it. In a second-preimage attack, the adversary starts with a specific message and tries to find a different message with the same digest. In a collision attack, the adversary is free to choose both messages. These are different problems, so a claim such as “the hash is secure” is incomplete unless the property and construction are named.

Blockchains use hashes primarily as commitments and identifiers. When a block header contains the hash of a previous header, any change to the earlier header changes the later reference. When a Merkle root commits to a set of transactions, an inclusion proof can demonstrate membership without sending the full transaction set. Proof-of-work systems can also use a hash function as part of the search puzzle.

Hashing does not provide confidentiality. If an attacker can guess the input space, they can hash guesses and compare digests. That is why password storage uses salts and deliberately expensive password-hashing constructions rather than a plain fast blockchain-style hash. Likewise, a hash does not authenticate an author: anyone who knows the input can compute the same digest. Authentication requires a keyed construction such as a MAC or a digital-signature scheme.

1. Define preimage, second-preimage, and collision resistance in your own words.

2. Why must collisions exist mathematically even for a secure fixed-length hash?

3. Explain how a block hash acts as a commitment to data without encrypting that data.

4. What additional property does a digital signature provide that a plain hash does not?

5. Why is a Merkle proof shorter than sending every transaction in a large block?

6. Correct this statement: “If data has been hashed, nobody can learn the original data.”