[Research] Post-Quantum Cryptography (PQC) Latency Benchmarks on OP Stack Architecture

Proposals 📃 Technical Proposals
1

Hi OP Stack Community and Core Devs,

As the transition toward Post-Quantum Cryptography (PQC) becomes an operational necessity across L1/L2/L3 infrastructures, the primary bottleneck remains execution overhead. Introducing lattice-based cryptography traditionally degrades sequencer throughput and expands block production times.

To address this, we have developed QUBEX Sentinel, a chain-agnostic PQC middleware engineered with a decoupled validation framework and high-concurrency Go execution engine, built to support L1, L2, and L3 networks.

We have successfully simulated a 100,000 transaction stress-test across 11 live network environments (including both L1s and various modular L2/L3 rollups) to measure real-world signing latency under NIST Level 5 boundaries.

Here are our verified performance metrics specifically for the OP Stack architecture:

Rank Network Type Avg Signing Latency Verification Status
1 Base Optimistic Rollup (OP Stack) 39,595 ns On-chain Ready
… [Other Tested Networks] L1 / L2 / L3 Frameworks [Data Kept Confidential] On-chain Ready
10 Optimism Optimistic Rollup (OP Stack) 246,632 ns On-chain Ready

Key Takeaway for OP Stack Builders:
Our benchmarks prove that under a decoupled architecture, Base can maintain sub-millisecond execution (39.5k ns) while running full quantum-resistant validation. However, the raw latency on the Optimism mainnet environment shows room for state-friction optimization (246.6k ns).

We have made our high-performance Go-based stress-testing suite public for independent verification. You can view the implementation and the terminal logs for these environments here:

We are currently looking to collaborate with OP Stack core developers, RaaS providers, and infrastructure researchers to run further stress-tests on the sequencer layer across L1/L2/L3 environments.

Would love to hear the community’s thoughts on optimizing PQC validation pathways before the quantum migration gap widens.

Best regards,
Spyridon Gagrinas
Founder, Qubex Sentinel

1 Like
2

I believe the quantum threat to ECDSA-based cryptography is one of the most critical yet underappreciated long-term risks for L2 ecosystems like Optimism.

The sub-millisecond latency results on Base (~39,595 ns) with ML-DSA are genuinely promising it shows PQC adoption at the sequencer level is not just theoretical anymore.

One suggestion: Alongside the technical benchmarks, publishing a clear Governance Migration Roadmap would be very helpful so that DAO communities can start building consensus around the transition before th @SpyrosGagr

3

Spot on. The quantum vulnerability of ECDSA is the elephant in the room for L2s.

The 39k ns benchmark on Base proves the OP Stack can natively handle NIST Level 5 without sequencer degradation—if validation is properly decoupled.

Technical readiness is only half the battle. Protocol consensus is the rest.

We are currently drafting the QUBEX Genesis Framework: a standardized migration blueprint for DAOs to upgrade sequencer security with zero downtime and zero user-side key rotation.

We’ll push the initial draft to the repo shortly. OP delegates and core researchers are welcome to review and contribute.

4

Update: Following up on the governance and protocol consensus discussion, the core team has expedited the initial release.

The QUBEX Genesis Framework (PQC Migration Blueprint) is now officially live in our repository.

It outlines the exact 3-phase integration pathway for the OP Stack to achieve NIST Level 5 security, ensuring Zero Network Downtime and Zero User UX Friction during the transition.

We invite all OP ecosystem delegates and protocol researchers to review the architecture here:

1 Like
5

Thanks for fast‑tracking and publishing the QUBEX Genesis Framework for OP Stack PQC migration. From a DAO risk‑management and governance perspective, I have a few follow‑up questions that would help delegates and risk stewards evaluate this more systematically:

  1. How do you propose Optimism governance should model quantum‑era signature compromise in the formal risk register over the next 5–10 years (e.g., likelihood × impact assumptions, indicative “time‑to‑compromise” ranges, and key dependency assumptions)?

  2. Within your three‑phase integration pathway, which specific checkpoints would you classify as governance‑gated because they materially change systemic risk for OP Stack chains (for example: new consensus or validation surface, changes to sequencer fault domains, or new failure modes that could affect liveness or safety guarantees)?

  3. For the “Zero Network Downtime” and “Zero UX Friction” guarantees, what minimum on‑chain and off‑chain telemetry would you recommend DAOs require to independently verify these claims during rollout (e.g., Chaos Engine metrics, latency distributions, error budgets, rollback / incident thresholds)?

  4. Finally, if OP Stack DAOs decide not to adopt a PQC migration blueprint in the near term, what worst‑case governance liability scenarios do you see emerging for tokenholders and delegates, given that quantum risk to ECDSA is now a known and discussed threat vector rather than an unknown unknown?

Clarifying these dimensions would make it much easier for governance participants to treat QUBEX not only as a technical pathway, but as a risk‑managed migration framework that can be embedded into Optimism’s broader security and upgrade process.

6

The core question for OP Stack governance is whether the PQC integration improves security, without affecting liveness or performance. Following NIST’s Post-Quantum Cryptography guidance ( Post-Quantum Cryptography | CSRC ), governance checkpoints should emphasise latency benchmarks, validator behaviour, rollback plans and independently verified telemetry.

As the quantum risks to ECDSA become more credible, PQC migration paths will become an important part of protocol risk management.

1 Like
7

@eliz Agree with your framing. The liveness-vs-security tradeoff is exactly the governance tension that needs to be resolved before any phase gate approval.

A few additions from a DAO risk-management perspective:

On latency benchmarks: Governance should require that telemetry be published on-chain (or in a verifiable off-chain registry) at each phase boundary not just as averages, but as P99 latency distributions under adversarial load. Average benchmarks can mask tail-risk that affects sequencer liveness.

On rollback plans: Each phase gate should have a pre-approved rollback condition encoded in governance a specific metric threshold (e.g., sequencer fault rate > X% over Y blocks) that automatically triggers a governance vote to pause or revert, reducing the dependence on discretionary human response during an incident.

On ECDSA quantum risk timeline: The shift from “unknown unknown” to “known discussed threat” does change the governance liability calculus for delegates. If this risk is on the register but no migration path is adopted, that becomes a documented decision which is fine, as long as it’s accompanied by a re-review trigger (e.g., annually, or upon any NIST update).