Cryptography & Key Sizing formula solvers

Computing & Chips
Computer ScienceCryptography & Key SizingSemiconductors & Chips
All 73 categories, 1,930 solvers

All 1,930 solvers →

Average Brute-Force Time

t=2 H−1Rt = \frac{2^{\,H-1}}{R}

Cryptography & Key SizingExpected time for an exhaustive search over a space of 2^H values at R guesses per second. The average is half the space, hence the H − 1. It assumes an OFFLINE attack on a fast hash with no rate limiting: the worst realistic case, and the only one worth sizing against.

Birthday Collision Probability (Exponential Approximation)

p≈1−e−n2/(2N)p \approx 1 - e^{-n^{2}/(2N)}

Cryptography & Key SizingChance that n values drawn independently from a space of N contain at least one repeat. This is the exponential APPROXIMATION, and it exists because the exact all-distinct product cannot be evaluated at the sizes cryptography works in: N = 2^128 would need 2^64 factors. It is excellent when n is far below N and it overstates the collision chance when n is a large fraction of N.

Elliptic Curve Security Strength

s=n2s = \frac{n}{2}

Cryptography & Key SizingClassical security strength of an elliptic curve group whose order is about n bits: half the key size, because the best known generic attack is Pollard's rho, which finishes in about the square root of the group order. P-256 gives 128 bits, P-384 gives 192, P-521 gives about 260.

Hash Collision Resistance (Birthday Bound)

n1/2≈2ln⁡2  ⋅  2 b/2n_{1/2} \approx \sqrt{2\ln 2}\;\cdot\;2^{\,b/2}

Cryptography & Key SizingNumber of messages that must be hashed before a collision becomes as likely as not, for a hash with a b-bit output. It is about 2^(b/2), and the constant sqrt(2 ln 2) = 1.1774 is the exact leading factor. NIST SP 800-107 Rev. 1 uses this to state that an approved hash function of b bits provides b/2 bits of collision resistance.

Key Space Size

N=bLN = b^{L}

Cryptography & Key SizingHow many distinct values exist when each of L positions can independently hold any of b symbols. A four-digit PIN is 10^4 = 10,000; a twelve-character password over the 95 printable ASCII characters is 95^12, about 5.4 × 10^23. This is a count of possibilities and nothing more: it is an upper bound on the work an exhaustive search needs, never an estimate of the work a real attack needs.

Key Stretching Work Factor

Δs=log⁡2niter\Delta s = \log_2 n_{\mathrm{iter}}

Cryptography & Key SizingBits of extra strength bought by making each guess expensive. Running a password through n iterations of a key derivation function multiplies an attacker's cost by n, which is log2(n) bits. The cost PARAMETER of bcrypt, scrypt and Argon2 is this number directly: a bcrypt cost of 12 means 2^12 = 4096 iterations and 12 added bits.

Passphrase Entropy (Diceware and Word Lists)

H=wlog⁡2DH = w\log_2 D

Cryptography & Key SizingEntropy in bits of a passphrase built by drawing w words uniformly at random from a list of D. Diceware's standard list holds 7776 words (6^5, one word per roll of five dice), so each word contributes exactly log2(7776) = 12.925 bits. The dice matter: rolling them is what makes the draw uniform, and choosing words that 'feel random' is not the same operation.

Password Entropy (Uniform Random Generation)

H=Llog⁡2bH = L\log_2 b

Cryptography & Key SizingEntropy in bits of a password GENERATED by picking L symbols uniformly at random from an alphabet of b. This page takes a length and an alphabet size and has nowhere to type a password, on purpose: entropy describes how a secret was produced, not what it looks like, so no calculation on the characters of a human-chosen password can recover it.

Reed-Solomon Correctable Errors

t=⌊n−k2⌋t = \left\lfloor \frac{n - k}{2} \right\rfloor

Cryptography & Key SizingSymbol errors a Reed-Solomon code RS(n, k) can correct: half the parity symbols, rounded down. Reed and Solomon's 1960 construction meets the Singleton bound exactly, so the n − k parity symbols do the most any code of that size could: they correct t errors at unknown positions, or twice as many erasures at known ones.

Truncated Hash Strength

scoll=b2spre=bs_{\mathrm{coll}} = \frac{b}{2} \qquad s_{\mathrm{pre}} = b

Cryptography & Key SizingSecurity a hash retains when its output is truncated to b bits, following NIST SP 800-107 Rev. 1: collision resistance falls to b/2 bits while preimage resistance stays at the full b. Truncation is permitted and normal (take the leftmost bits), but the two resistances shrink at different rates, and confusing them is how a digest ends up too short.

Unicity Distance

U=H(K)DU = \frac{H(K)}{D}

Cryptography & Key SizingShannon's 1949 figure for how much ciphertext a cryptanalyst needs before exactly one key remains consistent with it: the entropy of the key divided by the redundancy of the plaintext language. Below the unicity distance, several keys produce sensible-looking plaintexts and the message is genuinely ambiguous. It is a theoretical bound on INFORMATION available, not a statement about the work of finding the key.