Simulation-Secure OTMs

• Simulation-secure OTMs are cryptographic primitives that enable a receiver to access exactly one of two messages while ensuring that any adversary’s view is simulation-indistinguishable from an ideal protocol.
• They leverage quantum, hardware-based, and stateless token techniques to enforce irreversibility and limit leakage through carefully controlled single-query interfaces.
• Their design is foundational for one-time programs, quantum copy-protection, and composable cryptographic protocols, balancing rigorous security proofs with practical implementation challenges.

A simulation-secure one-time memory (OTM) is a cryptographic primitive enabling a receiver to learn exactly one of two messages while ensuring that the security of the scheme can be reduced to an ideal functionality via simulation. Simulation security requires that any adversary interacting with a real implementation is indistinguishable, up to negligible probability, from an adversary interacting with a simulated protocol that only accesses an ideal, single-query, minimal-leakage OTM interface. This notion is central to composable cryptographic design and forms a foundational building block for one-time programs, classical and quantum software copy-protection, and applications where “single-use” inaccessibility is required. There is extensive research establishing both quantum and hardware-enforced constructions, as well as impossibility results, for simulation-secure OTMs.

1. Ideal Functionality and Simulation Security

The ideal OTM functionality, denoted $\mathcal{F}_\mathrm{OTM}$, is defined as a two-party resource: the sender provides two messages $(s_0, s_1)$, and the receiver is permitted to submit a single selection bit $a\in\{0,1\}$, whereupon it receives $s_a$ and all information about the other bit is erased or remains inaccessible. All further attempts to query are met with a null value. The simulation-based security notion requires that for any adversary in the real-world model, a (possibly quantum or computationally unbounded) simulator exists that, given only single access to $\mathcal{F}_\mathrm{OTM}$, produces an output distribution indistinguishable from that of the real-world adversary, including any side information or additional classical output (Stambler, 27 Mar 2025, Stambler, 19 Jan 2026, Broadbent et al., 2018, Broadbent et al., 2015, Zhao et al., 2019).

In the universal composability (UC) framework, this demands indistinguishability even in arbitrary environments and under parallel composition. The formal definition is expressed as statistical or computational distance $\leq \operatorname{negl}(n)$ between the real and ideal output views for all $(s_0, s_1)$.

2. Core Construction Techniques

Simulation-secure OTM schemes rely on enforcing irreversibility and limited-extractability, most prominently leveraging quantum information-theoretic constraints or trusted tamper-resistance. Several key approaches:

• Quantum Random Access Codes (QRAC)-Based OTM: Prepares a tensor product state $$\rho = \bigotimes_{i=1}^n |\psi(c_{0,i},c_{1,i})\rangle$$ where each $|\psi(b_0, b_1)\rangle$ is a single-qubit encoding for bits $b_0, b_1$, typically chosen per the Ambainis–Nayak–Ta–Shma–Vazirani optimal $(s_0, s_1)$0 QRAC. Readout selects the $(s_0, s_1)$1 or $(s_0, s_1)$2 basis depending on which secret is being retrieved, and the outcome is decoded via a random linear code (Stambler, 27 Mar 2025).
• Wiesner-Style Conjugate Code OTM: Independently prepares single-qubit states $(s_0, s_1)$3 in $(s_0, s_1)$4 or $(s_0, s_1)$5 basis (Wiesner’s scheme), with classical hashes of the underlying string, and protects access to secrets using classical obfuscation and one-time pads. Correct retrieval is conditioned on measurement outcomes matching corresponding classical hashes for the selected basis (Stambler, 19 Jan 2026).
• Prepare-and-Measure with Stateless Tokens: Sender encodes secrets as quantum states correlated with classical data loaded into a stateless, tamper-resistant token. Receiver must measure qubits in either of two complementary bases and submit correct outcomes to the hardware token; only if classical checks pass is the corresponding secret revealed (Broadbent et al., 2018, Broadbent et al., 2015).
• TEE/Hardware-Based (Classical) OTM: Implements the ideal OTM by leveraging one-time use enforcement inside trusted execution environments. The TEE maintains a single-use flag and secret data, aborting all but the first query (Zhao et al., 2019).

3. Adversarial Models and Security Requirements

The OTM literature explores different adversary models according to available quantum or classical computational resources, adaptivity, and attack surfaces:

Model	Restrictions	Canonical Reference
C$(s_0, s_1)$6–GQNC$(s_0, s_1)$7	At most one non-adaptive quantum circuit (depth $(s_0, s_1)$8, geometrically $(s_0, s_1)$9-local), unlimited classical post-processing	(Stambler, 27 Mar 2025)
QROM QPT	Arbitrary quantum polynomial-time adversary, only classical queries to random oracle	(Stambler, 19 Jan 2026)
BPPQNC$a\in\{0,1\}$0 (Conjecture)	Quantum adversary with polynomial-depth circuits between classical random oracle queries	(Stambler, 19 Jan 2026)
QPT + Stateless Token	Unlimited quantum computation, classical queries to stateless hardware	(Broadbent et al., 2018, Broadbent et al., 2015)
Hardware (TEE)	Bounded physical tampering, assumes TEE is ideal and unforgeable	(Zhao et al., 2019)

A common feature is that the scheme is only required to be secure against a single, irrecoverable OTM query, and negligible leakage about the unchosen secret is tolerated.

4. Information-Theoretic and Computational Tools

Security proofs for simulation-secure OTMs rely on quantifying information leakage via rigorous entropy measures and bounding adversarial advantages through analytic techniques:

• Rényi Collision Entropy: Given a random variable $a\in\{0,1\}$1, $a\in\{0,1\}$2, capturing the effective guessing probability. Mutual information in collision entropy, $a\in\{0,1\}$3, serves as a metric for cumulative leakage during measurement (Stambler, 27 Mar 2025).
• Progress Bounds and Light-Cone Analysis: For constant-depth, geometrically-local quantum circuits, one decomposes the system into a “shell” plus independent blocks such that the quantum adversary’s measurements increase collision mutual information only marginally per block. The total leakage scales as $a\in\{0,1\}$4, with $a\in\{0,1\}$5 the system size (Stambler, 27 Mar 2025).
• Min-Entropy Bounds and Uncertainty Relations: For quantum adversaries, impossibility of extracting both secrets is witnessed by strong lower bounds on the probability of simultaneously guessing classical labels prepared in conjugate bases, e.g., the “sequential conjugate-coding bound” for POVMs: success in the $a\in\{0,1\}$6 basis forces negligible guessing probability in the $a\in\{0,1\}$7 basis, formalized as $a\in\{0,1\}$8 (Stambler, 19 Jan 2026).
• Semidefinite Programming (SDP): The ultimate cheating probability for stateless-token schemes, i.e., the chance to obtain both $a\in\{0,1\}$9 and $s_a$0, is characterized precisely via SDP. For linear probability of queries $s_a$1 with $s_a$2, $s_a$3 (Broadbent et al., 2018).
• Distributional VBB Obfuscation and LPN: Computational constructions (not information-theoretic) invoke distributional VBB (virtual black-box) obfuscation of conjunctions, constructible from the Learning Parity with Noise (LPN) assumption (Stambler, 19 Jan 2026).

5. Efficiency, Practicality, and Parameter Choices

Realizing simulation-secure OTMs in practice involves nontrivial trade-offs in efficiency, state preparation, and error rates:

• Quantum Code Approaches: The security parameter is the number of qubits $s_a$4; statistical cheating probability decreases exponentially in $s_a$5. However, decoding random linear codes required in QRAC-based schemes is exponential-time in the worst case (Stambler, 27 Mar 2025). For QRAC-based OTMs, correct decoding occurs except with probability $s_a$6, with message length $s_a$7 for any rate $s_a$8.
• Prepare-and-Measure: These schemes require only single-qubit state preparations and projective measurements in $s_a$9 or $\mathcal{F}_\mathrm{OTM}$0 basis, implementable within current experimental capacities (e.g., NV centers, photonic systems), as no long-term quantum memory or entanglement is needed (Broadbent et al., 2018, Broadbent et al., 2015).
• Hardware Token and TEE: The trusted execution variant offers scalability limited by hardware and system startup costs. TXT-only variants optimize for small-sender/large-receiver settings, while garbled-circuit based variants achieve best performance for large-sender/small-receiver domains (Zhao et al., 2019).
• Assumptions and Open Questions: Quantum OTMs often rely on tamper-proof hardware or physical principles (no-cloning, measurement disturbance). TEE-based schemes presume unbreakable sealing/integrity and lack of side-channel attacks. Fully classical, information-theoretic OTM in the standard model remains open.

6. Extensions, Limitations, and Impossibility Results

The limitation and strength of simulation-secure OTMs are inherently model-dependent.

Impossibility Results:

• OTMs are impossible to achieve information-theoretically in the plain classical or quantum standard model without additional assumptions (Broadbent et al., 2018, Broadbent et al., 2015).
• Security collapses if hardware tokens are subject to quantum superposition queries, allowing adversaries to extract both secrets by quantum rewinding and uncomputation, analogous to known attacks on classical one-time programs (Broadbent et al., 2018).
• Measure-and-access constructions require exponentially many valid keys per secret; otherwise, the adversary can guess all keys with non-negligible probability.

Extensions and Conjectures:

• Security can extend to adaptive, polynomial-depth quantum adversaries (BPPQNC$\mathcal{F}_\mathrm{OTM}$1), assuming conjectured lifting theorems in the quantum random oracle model. The predicate "retrieve both secrets" remains approximately as hard as for classical-query-only adversaries under these models (Stambler, 19 Jan 2026).
• Tolerating polynomially many adaptive token queries without sacrificing simulation security is unresolved for quantum/hardware models (Broadbent et al., 2018).

Parameter Regimes:

Construction Type	Message Length	Adversarial Bound	Complexity
QRAC-based Quantum	$\mathcal{F}_\mathrm{OTM}$2	non-adaptive, const-depth, local	Exponential-time decoding
Wiesner + Obfuscation	poly$\mathcal{F}_\mathrm{OTM}$3	quantum polytime + ROM (classical queries)	Polytime, assuming LPN
Stateless Token	$\mathcal{F}_\mathrm{OTM}$4	$\mathcal{F}_\mathrm{OTM}$5 queries	Polytime
TEE/Hardware	Arbitrary	Hardware-limited	Practical, TEE-dependent

7. Related Applications and Research Directions

Simulation-secure OTMs are universal for constructing one-time programs and form a key primitive in quantum cryptographic design. Recent works provide concrete frameworks with composable security, efficient implementations in hybrid models, and propose new bounds on quantum adversarial limitations (Stambler, 27 Mar 2025, Stambler, 19 Jan 2026). The field remains open with respect to achieving efficient, polynomial-time, unconditional security exclusively via quantum or classical computation with minimal assumptions, and research continues on adapting OTM notions for noise-tolerant, reusable, and composition-secure extensions. The relationship between entropy-based leakage, depth-limited quantum adversaries, and practical enforceability remains a vibrant area for further investigation.

References:

(Stambler, 27 Mar 2025, Stambler, 19 Jan 2026, Broadbent et al., 2018, Broadbent et al., 2015, Zhao et al., 2019)