Public-Key Cryptography and Digital Signatures
Last updated: June 10, 2026
Why It Matters
Blockchains have no login page. There is no “forgot password,” no fraud department, no one to call. The entire notion of ownership — who can spend a coin, who controls a contract — reduces to one mechanism: can you produce a valid signature with the right secret key?
That’s why the casebook’s warning (Gorbunov) deserves its bluntness: private keys are money. Most “blockchain hacks” aren’t broken cryptography — they’re buggy wallets, leaked keys, and people rolling their own crypto. The math has held; the key management around it is where funds die.
How It Works
Beginner
Imagine a magic pen and a public autograph card. Anyone can compare a signature against your card and confirm it’s genuinely yours — but nobody can forge the pen’s signature, and your signature on one document can’t be peeled off and stuck onto another. The pen is your secret key (never share it); the card is your public key (share it freely). Spending crypto is signing a note that says “I authorize this payment” with your pen.
Intermediate
A signature scheme has three parts, per the casebook:
- Generate — produce a keypair: a public key (your identifier on the network — addresses derive from it) and a secret key.
- Sign — combine the secret key with a specific message (the transaction) to produce a signature.
- Verify — anyone with the public key, message, and signature can check validity.
Two guarantees follow: consent (only the secret-key holder could have produced it) and integrity (the signature is message-specific — alter the payee or amount and verification fails). This is what secures the UTXO Model: every input must carry a valid signature from the keys its locking script demands.
Builder
Bitcoin uses ECDSA (and, since Taproot, Schnorr) over secp256k1; Ethereum uses ECDSA/secp256k1 with Keccak-derived addresses. Schnorr’s linearity enables key and signature aggregation (MuSig) — n-of-n multisig that looks like a single signature on-chain. What’s actually signed is a hash of the transaction (sighash modes select which parts), which is why hash functions and signatures are inseparable in practice. Standing advice from the casebook stands verbatim: never roll your own crypto.
Examples
- Bitcoin — ECDSA/Schnorr signatures authorizing every UTXO spend.
- Ethereum — ECDSA account signatures; contract wallets add programmable policies.
- Multisig custody — m-of-n signatures for treasuries and exchanges.
- Everywhere else — TLS, SSH, code signing: the same primitive securing the internet.
Tradeoffs
Strengths
- Self-sovereign ownership — no registrar or intermediary stands between a keyholder and their assets.
- Unforgeable, message-specific consent — the two properties that make permissionless value transfer possible.
- Verifiable by anyone — public verification is what lets every node check every transaction.
Limitations
- Keys are money — loss is irrecoverable, theft is final; the human layer is the weakest link.
- No revocation — a compromised key can’t be “cancelled”; funds must be raced to a new key.
- Quantum horizon — large-scale quantum computers would break ECDSA; post-quantum migration is an open, slow-moving question for every chain.
Related Concepts
- UTXO Model — Where signatures meet spendable coins
- Double-Spend Problem — Consent alone isn’t enough; ordering matters too
- Smart Contracts — Programmable spending conditions beyond a single signature
- Hash Functions — The companion primitive
Sources & Last Updated
- MIT BLC Module 2: Maintaining Blockchain Integrity (primary source; Gorbunov lecture)
- Vault note: Public-Key Cryptography and Digital Signatures (M2 cluster)
Last updated: June 10, 2026