Research
SpexPay: A Privacy-Preserving Pay-As-You-Go System for Dynamic Spectrum Sharing
Dynamic spectrum sharing today runs through centralized Spectrum Access Systems (SAS) run by commercial administrators such as Google, CommScope, Federated Wireless and Sony.

- arXiv
- 2607.17218
- Published
- 2026-07-19
- Authors
- Mohaimin Al Barat, Hexuan Yu, Shaoyu Li, Yang Xiao, Yi Shi, Eric W. Burger, Y. Thomas Hou, Wenjing Lou
AI summary
Overview
Research area: Cryptography and security applied to dynamic spectrum sharing (DSS) — specifically privacy-preserving, usage-metered payments for shared radio spectrum.
Technical level: Advanced. The paper assumes familiarity with BBS+ signatures, Schnorr-style zero-knowledge proofs, the Fiat–Shamir transformation, bilinear pairings, hash chains, and Solidity smart contracts.
Scope: A full system design, security analysis, and prototype evaluation of SpexPay, a privacy-preserving pay-as-you-go (PAYG) spectrum-access framework built on verifiable credentials plus an on-chain escrow contract.
What This Paper Is About
Dynamic spectrum sharing today runs through centralized Spectrum Access Systems (SAS) run by commercial administrators such as Google, CommScope, Federated Wireless and Sony. Every access request leaks location, time, frequency, bandwidth and dwell time to that commercial middleman, which over time allows an operator's mobility patterns or mission profiles to be reconstructed. The paper asks whether a user can pay for spectrum strictly by how much they actually consume, without ever revealing who they are — while still leaving regulators a way to audit and settle disputes. SpexPay answers this by combining unlinkable two-tier credentials, hash-chained usage "heartbeats," and a smart contract that pays a Band Manager only for cryptographically proven usage.
Key Contributions
-
A PAYG model for spectrum sharing. The authors introduce SpexPay, which they describe as the first PAYG system for next-generation spectrum economies combining fine-grained, usage-based micro-payments with the regulatory functions of incumbent paradigms.
-
Privacy-preserving access control and micro-payments. A two-tiered spectrum credential system lets user devices authenticate for spectrum usage without revealing real identities to Band Managers. The hidden device identifier must be provably identical across both credentials yet never become a cross-session correlation handle.
-
Verifiable usage metering without per-interval signatures. Instead of signing each heartbeat, devices emit precomputed forward-secure hash-chain values, which the Band Manager checks by rewinding the chain, and which the contract replays to authorize payment.
-
A prototype and evaluation. The implementation measures ≈150 ms for credential issuance, proof generation and verification, ≈603K gas (≈$0.9) for smart-contract settlement, and reports a Raspberry Pi 5 evaluation of user-side cryptographic operations to assess suitability for edge-class hardware. Code and artifacts are public at https://github.com/iambarat/SpexPay.
Main Findings
- Low end-to-end latency: The prototype demonstrates ≈150 ms total computation time for credential issuance, proof generation and verification, which the authors argue is practical for real-world DSS deployment.
- Modest on-chain cost: Settlement costs ≈603K gas, quoted as ≈$0.9, because only constant-size values (session ID, heartbeat index, one hash) are committed on-chain.
- Edge feasibility: User-side cryptographic operations were evaluated on a Raspberry Pi 5, which the authors state confirms scalability to edge-class hardware. Per-operation timings are not reported in the available excerpt.
- Identity is decoupled from settlement: Identity-bound credentials stay off-chain in unlinkable BBS+ selective-disclosure presentations, while the chain records only ephemeral session identifiers and hash-chain commitments. This is the mechanism that reconciles privacy with auditability.
- Three formally stated security properties: Theorems are given for unforgeability, unlinkability, and fair settlement, resting on the q-SDH assumption, discrete-log hardness in the Pedersen commitment group, collision and preimage resistance of the hash function, the random oracle model for Fiat–Shamir challenges, and standard blockchain consensus properties. Collusion resistance is named as the fourth core property, but its theorem appears beyond the truncation point of the provided text.
- Replay and over-claiming are blocked at the contract level: Each heartbeat can be claimed once, the contract deterministically recomputes hash-chain links between the last claimed index and the new one to rule out skipped, injected or inflated intervals, and each hash is domain-separated to (session ID, index) to prevent cross-session reuse.
- Off-chain audit registry: The commercial (Type-B) issuer maps session IDs to device IDs privately, so a regulator can resolve disputes using on-chain data without the chain or the Band Manager ever learning a real identity. A separate session revocation list is shared with Band Managers.
- Comparison to prior work is qualitative: The paper contrasts SpexPay against obfuscation-based approaches (Bahrak et al., Clark et al.), TrustSAS, Pri-Share, Coconut, Zerocash, PrivPay and Monero, but reports no head-to-head numeric benchmarks against them in the available text.
Methodology in Plain English
The design mirrors how regulated spectrum access already works, so no new actors are invented. Four participants exist: a regulatory issuer (e.g., the FCC), a commercial issuer (e.g., a mobile network operator that a device belongs to), a Band Manager that grants and monitors access, and the user device itself.
Each device gets two credentials signed with BBS+ signatures. A long-term regulatory credential certifies eligibility (access tier such as PAL or GAA, device class, expiry). A short-lived commercial credential covers pricing, usage limit and a unique session ID. Both embed the same device ID as a hidden attribute. When the device wants spectrum, it generates a zero-knowledge presentation showing both credentials are valid and that the hidden IDs match — without revealing the ID or any attribute it chooses not to disclose.
Money works through escrow. Before each session, the commercial issuer deposits tokens into a smart contract and publishes session parameters: session ID, the root of a hash chain, the heartbeat limit L, a per-heartbeat unit price, an expiry time, and the Band Manager's address. Because the chain runs backwards from a random final value to the root, a device can reveal one hash per interval but can never precompute future values — so neither the user nor the Band Manager can fabricate usage. During the session the device sends a heartbeat (session ID, index, hash) roughly every 5 minutes. The Band Manager verifies by hashing forward from the received value and comparing against the last stored hash, and can authenticate skipped indices by iterating backward.
To get paid, the Band Manager submits the latest valid heartbeat to the contract. The contract replays the chain from the last claimed index, computes the delta, multiplies by the unit price, checks the payout does not exceed the deposit, updates state, and transfers. When the session expires, the commercial issuer closes it and any unspent deposit is refunded. The threat model treats Band Managers and users as untrusted but assumes issuers are honest within their trust boundaries and that the chain is tamper-resistant.
Why This Matters
Impact on research. The paper argues a specific gap: existing spectrum-privacy work improves allocation but lacks economic enforcement, while anonymous-credential and private-payment systems are general-purpose and do not encode access eligibility or bind payment to externally verifiable service consumption. SpexPay claims to be the first to combine unlinkable access authentication, verifiable usage metering, and on-chain non-repudiable settlement — a template likely to be reused wherever a regulated service must be metered continuously rather than paid for once.
Real-world applications:
- Military and federal incumbents such as shipborne radar, where the paper argues privacy is legally mandated rather than merely desired, since aggregated grant requests reveal mobility trajectories and operational rhythms.
- Commercial mobile network operators running CBRS (3.55–3.70 GHz) and other shared bands who want non-repudiable billing evidence while shedding metadata-storage liability.
- Radio astronomy stations and Programme Making and Special Events (PMSE) operators under Europe's Licensed Shared Access framework and national local-licensing regimes, which the paper says expose comparable metadata.
- Cellular, IoT and sensing deployments where ephemeral, session-based access is a better fit than long-term leases.
Industry relevance. The regulator can mandate the privacy layer as a condition of SAS certification, giving it system-wide leverage without requiring any new market participant. A ≈$0.9 settlement cost and 150 ms latency sit within the operational envelope of a real registry, and the Raspberry Pi 5 result suggests commodity edge hardware suffices.
Future Directions
- Finishing the collusion-resistance argument. Collusion resistance is listed as one of four core properties but its theorem falls outside the excerpt; the precise adversarial model — presumably commercial issuer colluding with Band Manager, or user with Band Manager — remains to be read in full.
- Spectrum-availability discovery. The authors explicitly defer the question of how a device learns what spectrum is available in its zone, calling it orthogonal and assuming it is already obtained. Integrating that query without reintroducing a location leak is an open design problem.
- Broader evaluation. The reported results are a single prototype. Independent comparison against TrustSAS, Pri-Share or general anonymous-payment systems on latency, gas and privacy guarantees would strengthen the case.
- Deployment and governance questions. Mapping the two-tier issuer model onto real regulators and MNOs — including who operates the private audit registry, how revocation propagates, and what dispute-resolution procedure sits above the contract — is left as policy and operational work.
Target Audience
Researchers and graduate students in applied cryptography, blockchain systems, and wireless spectrum policy; engineers designing SAS or CBRS-adjacent infrastructure; and regulators or standards participants evaluating privacy requirements for shared-spectrum marketplaces. The paper assumes comfort with pairing-based signatures and zero-knowledge proof protocols, so readers without that background will need supporting material on BBS+ and Schnorr proofs.
Authors’ abstract
Dynamic Spectrum Sharing (DSS) is a cornerstone of next-generation wireless systems, yet existing solutions such as Spectrum Access Systems (SAS) rely on centralized administrators that expose sensitive operational metadata and lack cryptographic transaction accountability. Though SAS administrators, such as Google, have introduced pay-as-you-go pricing models, these approaches still face significant privacy and accountability challenges as DSS evolves toward a more open and…
The full abstract on arXiv