Empirical/QM: minimal reusable security-protocol interface #
Category: 3-Local (QM-validity scaffolding, no CSD ontology).
A small, deliberately minimal set of definitions giving the cryptographic
tranches (Crypto/E91KeyRate.lean, Crypto/QuantumMoney.lean) a shared
vocabulary for "an adversary's advantage is bounded" and "a real protocol
extracts a positive secret-key rate".
Honest scope. These are formal stand-ins for the protocol / security
framing, not a full composable security model. There is no
universally-composable (UC) / abstract-cryptography (AC) simulator, no
adversarial-view (AVS) distinguisher, no finite-key smoothing, and no
information-theoretic ε-secrecy proof here. SecurityBound records a single
real number ε with the "adversary cannot do better than ε" reading;
IdealQKD records only a key length plus a Prop-valued secrecy stand-in;
RealProtocol.secure is the bare "positive key rate ⟹ extractable key"
predicate. The deeper notions (composability, min-entropy accounting,
finite-key corrections) are named here and left out of scope; this file
provides only the reusable interface that the key-rate and forgery statements
instantiate.
Field consumption (honest, future-facing). SecurityBound.ε is consumed
now by Crypto/E91KeyRate.e91_eavesdropper_advantage, which proves the
LHV/eavesdropper advantage is bounded by it (grounded in lhvCHSH_abs_le_two),
and is the slot the quantum-money tranche will reuse as forgery probability.
IdealQKD.idealSecret is a placeholder carried for the QKD finite-key /
composable-secrecy tranche that will consume it; at this interface layer it is
deliberately abstract.
An adversary's success / distinguishing bound: the single number ε,
constrained to a probability 0 ≤ ε ≤ 1, with the reading "no adversary
succeeds (forges / distinguishes / learns the key) with probability greater
than ε". Reused by both the QKD tranche (eavesdropper information) and the
quantum-money tranche (forgery probability). Minimal by design: a richer
security object (min-entropy, smoothing parameters) is out of scope here.
- ε : ℝ
The adversary-advantage bound.
Advantages are nonnegative.
Advantages are probabilities, hence
≤ 1.
Instances For
The ideal QKD functionality, as a minimal stand-in: a key length
keyLength (in bits) together with a Prop-valued placeholder idealSecret
standing for "the ideal key is uniform and independent of any adversary".
This is explicitly not a uniform-distribution object or a composable ideal
functionality; it is the smallest datum the real-vs-ideal predicate below
needs to mention.
- keyLength : ℕ
The ideal key length in bits.
- idealSecret : Prop
Stand-in proposition for "the ideal key is uniform and secret". Carried abstractly; a concrete uniform-distribution / min-entropy witness is out of scope at this interface layer.
Instances For
A real protocol producing a key: an asymptotic secret-key rate
keyRate (secret bits per signal) together with the SecurityBound it
achieves against the eavesdropper.
- keyRate : ℝ
Asymptotic secret bits extracted per transmitted signal.
- security : SecurityBound
The adversary-advantage bound the protocol guarantees.
Instances For
A real protocol is secure when it extracts a strictly positive
secret-key rate: a positive asymptotic rate means a genuine secret key can be
distilled (for keyRate ≤ 0 no key is extractable).
Instances For
P emulates the ideal functionality I over n signals (asymptotic
stand-in): the ideal keyLength bits are covered by the n · keyRate secret
bits the real protocol produces. This is the single consuming use of
IdealQKD; it is an asymptotic accounting identity, not a composable-security
emulation.
Instances For
A secure protocol (positive key rate) emulates any ideal key length
once enough signals are sent: the asymptotic rate amortises any fixed key
length. Binds secure, emulates, RealProtocol, and IdealQKD into one
statement, so each definition has a genuine consumer.