Every time you open a website, send a message or install an update, cryptography is protecting you.
All of it rests on a small number of bets about what computers can't do. These maps show those bets and everything
built on top of them, so you can see what would break if one of them turned out to be wrong.
Last reviewed 10 October 2026. New research and standards are checked weekly.
1. Cryptography rests on bets
Most cryptography works because some math problems are easy in one direction and extremely hard in the other.
Multiplying two huge prime numbers takes a computer a split second. Going backwards, from the product to the two primes,
is believed to be out of reach of every computer on Earth at the sizes used in practice.
Nobody has proved that these problems are hard. People trust them because the best mathematicians have attacked them
for decades without finding a shortcut. The map calls them hardness assumptions. If one falls, everything built on it
falls too.
A few things need no such bet at all. They sit in the box called Unconditional: they stay secure even against
unlimited computing power, usually at a practical cost. A one-time pad, for example, needs a key as long as the message.
2. How a hard puzzle becomes a lock
A hard math problem is only useful if someone can do what everyone else can't. The same few tricks come back on
almost every page:
A secret starting point. Pick a secret number and publish where it lands after an easy step that can't be
undone, such as jumping around an elliptic curve. The landing spot is your public key, the secret is your private key,
and a signature proves you know the secret without revealing it.
A secret shortcut. Build a puzzle that is hard to undo for everyone except you, because you made it with
hidden information, like the two primes behind an RSA number. Anyone can lock a message with the public version. Only
you can unlock it. Cryptographers call the hidden information a trapdoor.
Noise as a disguise. Some data looks like random junk unless you know a secret: the output of AES, or
equations with small errors added (LWE, LPN, codes). Add that junk to a message like a one-time pad. Only someone who
can recompute it can peel it off.
Mixing secrets in public. Two people each mix a private ingredient into a public message, in a way that
lets both mixtures combine into the same result while an eavesdropper can't unmix them. That is Diffie-Hellman key
exchange.
Unforgeable fingerprints. If nobody can find two inputs with the same hash, a short hash can stand in for a
whole document, database or set of keys: signing, storing or publishing the hash fixes the whole thing.
On every box page, the "In plain words" box has a short The idea paragraph showing which trick that box uses,
or how it works inside.
3. The building blocks
On top of the bets sit a handful of tools. The small colored label above each box name says which kind it is:
Hash functions make a short fingerprint of any data. They check that downloads weren't tampered with and protect stored passwords.
Symmetric encryption uses one shared key to lock and unlock. It is fast and protects almost all data once a connection is set up.
Key exchange lets two strangers agree on a shared key over a network anyone can listen to. That key then feeds symmetric encryption.
Public-key encryption lets anyone lock a message with your public key, while only your private key unlocks it.
Digital signatures prove who wrote something and that it wasn't changed. Software updates and website certificates rely on them.
MACs are like signatures made with a shared key: they catch tampering in transit.
Secure computation lets several parties compute a result together without showing each other their inputs.
At the bottom of the general map are protocols people use every day: TLS (the padlock in your browser), Signal, SSH
and WireGuard. Each combines several building blocks, so it depends on all of their bets at once.
4. How to read an arrow
An arrow from A to B means: if A breaks, B breaks with it. Follow the arrows downward to see the fallout.
For example, a large quantum computer running Shor's algorithm[1] would solve factoring. That breaks the RSA problem,
which breaks RSA encryption and RSA signatures. On the map that is a chain of three arrows.
An arrow that ends on the edge of a big box means everything inside that box depends on it. A purple AND gate means
the box below only breaks if all of its inputs break: that is how hybrid schemes protect themselves.
The map is about the math. Real systems can also fail on their own, through a software bug, a stolen key or a bad
random number generator, without any arrow being involved.
5. Shapes, colors and tags
darkFoundational assumption. The deepest bets, such as "one-way functions exist" and "good randomness". Nearly everything hangs off them.
roundedHardness assumption. A math problem believed to be hard, like factoring.
dashedTrust assumption. A bet about people or setup rather than math, like "more than half of the participants are honest".
squareBuilding block or scheme. Something built from the assumptions above it, like a cipher or a signature scheme.
boldProtocol or component. Something people use directly.
✗ redHistorically broken. Once trusted, now broken. It stays on the map as a warning.
──▶If the source breaks, the target breaks.
━━▶What Shor's algorithm on a large quantum computer would break.
ROM GGM CRHSmall orange tags: the security proof leans on a simplification or an extra condition. Each box page explains its tags.
FSMade non-interactive with Fiat-Shamir: it falls with the Fiat-Shamir box, even though no arrow is drawn.
KEThis assumption is strong enough to give key exchange on its own, which needs more than one-way functions.
TDThis assumption gives trapdoor functions on its own: a function anyone can compute and only a secret can undo. It falls with the trapdoor box. Schemes built on it need not use a trapdoor themselves: ElGamal doesn't, yet decisional DH carries the tag.
6. The quantum question
A large enough quantum computer could run Shor's algorithm, which breaks factoring and discrete logarithms. That covers
RSA, elliptic curves and pairings: most of the public-key cryptography in use today. The red arrows on the map show
exactly what would fall. Nobody has built such a machine yet, which is why the map lists "no large-scale quantum
computer" as an assumption.
Hash functions and symmetric ciphers only face Grover's algorithm[2], a much smaller speedup that longer keys absorb.
Replacements for public-key cryptography, called post-quantum cryptography, rest on different bets: lattices,
error-correcting codes, hash functions and others. NIST published the first standards in 2024 (ML-KEM, ML-DSA and
SLH-DSA)[3, 4, 5]. Many systems now run a classical and a post-quantum scheme side by side, so an attacker would have to break both.
7. The Ethereum map
The Ethereum map applies the same idea to one system. The top half, Live on mainnet today,
is what protects Ethereum right now. The bottom half, Not live yet, holds planned upgrades in rough time order,
and each one says which assumptions it would rest on.
dottedPlanned, not live yet. Glamsterdam and Hegotá are the next named upgrades. I*, J* and K* are later forks on the Ethereum Foundation's draft roadmap, the strawmap.
···▷A roadmap step: one item leads to the next. It is not a dependency, so nothing falls along it.
Arrows into Ethereum components are colored by where they come from: red for elliptic curves (which a quantum computer
would break), green for hashes (which it wouldn't), and amber for trust assumptions.
8. Where to start
Every box opens a page that begins with a short explanation in plain words. Tap or hover any word with a dotted
underline for a quick definition, or browse the glossary. Some good first boxes:
TLS 1.3The "S" in HTTPS: it sets up a private, authenticated connection to a website.
Collision resistanceNobody can find any two different inputs with the same fingerprint.
Integer factorizationMultiplying two huge prime numbers is easy. Going backwards, from the product to the two primes, is believed to be extremely hard.
No large-scale quantum computerA large enough quantum computer would easily solve the math behind today's digital padlocks (RSA and elliptic curves). Nobody has built one yet.
One-time padLock a message with a truly random key as long as the message, and never reuse it. That is mathematically unbreakable, but you need to hand over a mountain of key in advance.
On the Ethereum map:
Accounts (EOAs)Ordinary accounts controlled by a single private key.
BLS signaturesShort signatures that can be merged: thousands of validator votes become a single signature.
KZG commitmentsSeals a whole blob of rollup data into one tiny value, with tiny proofs for any piece.
P.W. Shor. Algorithms for quantum computation: discrete logarithms and factoring. Proceedings 35th Annual Symposium on Foundations of Computer Science, 2002. doi:10.1109/sfcs.1994.365700
Lov K. Grover. A fast quantum mechanical algorithm for database search. Proceedings of the twenty-eighth annual ACM symposium on Theory of computing - STOC '96, 1996. doi:10.1145/237814.237866