Signing Latency Audit

Know your actual p50 and p99 signing latency under realistic load, compared against every alternative you might want to migrate to.

Most teams don't know their signing p99 until they're debugging a latency incident. By then they're under pressure, making architecture choices without data. This audit gives you the data before that moment.

For reference: the ZeroCopy TEE signing pipeline clocks 42µs, which is 3,095× faster than AWS KMS on the same workload. I'll tell you where your stack lands on that same benchmark.

$1,500 deposit to reserve your slot. Remainder due on delivery. Non-refundable deposit, full report guaranteed or you don't owe the balance.

42µs TEE signing 3,095× AWS KMS 1,245 tests green AI orchestration: Letta + BAML

Reserve your slot — $1,500 deposit Discuss scope first (free 20-min)

What you get.

Signing path instrumentation on your prod or staging environment

I add OpenTelemetry spans to your signing path (or review existing instrumentation if you have it). Every key operation gets a label: HSM call, key derivation, serialization, network round-trip. You see where the time actually goes.

p50/p99 measurement under realistic load

Not synthetic benchmarks. I replay your actual transaction rate (or a reasonable approximation using your stated TPS) against the instrumented path. The numbers in the report reflect what your system does under load, not what a demo script achieves.

Comparison vs TEE, CloudHSM, AWS KMS, and colo HSM

I run the same workload against reference implementations (or use measured baselines where I can't instrument directly) and give you a comparison table. You'll know exactly how much headroom you're leaving on the table, or whether your current stack is already competitive.

Threat-model alignment review

Latency isn't the only axis. I check whether your signing path handles key compromise, replay attacks, and side-channel exposure appropriately given your threat model. A 42µs signing path that leaks the key material is worse than a 10ms path that doesn't.

Remediation roadmap + 1-hour walkthrough

Written roadmap with specific changes ranked by latency impact. Followed by a 60-minute call where I walk your team through the findings and implementation options.

What this is not.

Not a full security audit. I look at the signing path specifically. I'll flag obvious adjacent issues I notice, but this is a latency and threat-model review, not a comprehensive penetration test or code security audit.

Not a migration project. The audit tells you what to change and why. Executing the migration is a separate engagement. If you want to move from KMS to TEE after the audit, that's a scoped project we can discuss.

Not applicable to software signing (code signing, container signing). This audit covers transaction signing, message authentication, and key management for financial or high-assurance systems. Package signing and CI/CD code signing pipelines are a different workload.

Not useful if you don't control your signing stack. If you're fully delegating signing to a custody provider and have no visibility into their implementation, there's nothing to instrument. The free 20-minute diagnostic will clarify whether this applies.

7-day process.

Day 1

Kickoff

You share stack access and existing metrics (if any). I review your signing architecture and instrument the key path with OpenTelemetry spans. We agree on the load profile for the benchmark.

Day 2-4

Measurement

Load replay against instrumented path. I collect p50, p90, and p99 at your stated TPS and at 2x and 5x burst. I run the comparison workload against reference implementations and build the comparison table.

Day 5-6

Threat model + roadmap

I review your signing path against your threat model. I draft the remediation items, rank them by latency impact and risk, and cross-check them against your operational constraints.

Day 7

Report + walkthrough

Written report delivered (PDF + markdown). 60-minute walkthrough call with your team. Balance due on delivery. Instrumentation code (if I wrote it) is yours to keep regardless.

Questions.

What's the refund policy on the $1,500 deposit?

The deposit is non-refundable once I've started work (Day 1 kickoff). Before kickoff, I'll refund in full if you cancel at least 48 hours before the scheduled start. If I fail to deliver the report by Day 7, you owe nothing for the balance. The deposit still stands because I've already spent time on kickoff and instrumentation.

Who owns the report?

You do. Full IP transfer on final payment. I retain the right to reference the engagement type in aggregate, but will not disclose client name, findings, or benchmark numbers without written permission. Any instrumentation code I write during the engagement is also yours.

Can I add scope mid-audit?

No. The 7-day window is fixed. If you have additional signing paths or systems you want benchmarked, we schedule a second engagement after the first report delivers. Returning-client discount of 15% applies.

Know your number.

"Our signing is fast enough" is not a number. A p99 of 280ms at peak load is a number. You need the number before you can decide whether to fix it.

Reserve your slot — $1,500 deposit Scope question first (free 20-min)