DEV Community

Skeletor-Pirate
Skeletor-Pirate

Posted on

Why We Need to Schedule Post-Quantum Cryptography Around the Weather

Why We Need to Schedule Post-Quantum Cryptography Around the Weather

The security industry is undergoing one of the largest migrations in history: moving from classical cryptography (RSA/ECC) to Post-Quantum Cryptography (PQC).

While algorithms like ML-KEM (FIPS 203) are quantum-resistant, they have a hidden cost: they are CPU-intensive. Compared to classical algorithms, key generation, signing, and verification consume significantly more compute.

More compute = more electricity. For bulk operations (signing millions of docs, encrypting petabytes of backups), this spikes carbon footprint — bad for net-zero goals.

What if we schedule these workloads to run only when the grid is cleanest?

Enter Quant Harvest.

What is Quant Harvest?

Quant Harvest is an open-source Go control plane — a carbon-aware deferred-execution queue for PQC workloads.

Instead of executing heavy crypto immediately, it buffers workloads and monitors grid carbon intensity (gCO2e/kWh). Dirty grid (fossil) → pause queue. Clean grid (solar/wind) → resume and harvest.

                  ┌──────────────────────┐
                  │   Workload Enqueued  │
                  └──────────┬───────────┘
                             ▼
                  ┌──────────────────────┐
                  │    Task Buffered     │
                  └──────────┬───────────┘
                             ▼
               No  /───────────────────\
             ┌────/  Grid Clean enough? \
             │    \   (Under threshold) /
             │     \───────────────────/
             ▼               │ Yes
     ┌──────────────┐        ▼
     │ Queue Paused │ ┌──────────────┐
     └──────────────┘ │ Execute Task │
                      └──────────────┘
Enter fullscreen mode Exit fullscreen mode

Try it live: https://quant-harvest.onrender.com

curl -X POST https://quant-harvest.onrender.com/api/tasks \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: demo-123" \
  -d '{"name":"pqc-batch-signing","payload":"ml-kem-768"}'

# Watch /api/status — when carbon >200 it defers, when <200 it runs
curl https://quant-harvest.onrender.com/api/status | jq
Enter fullscreen mode Exit fullscreen mode

Architecture

  • Go stdlib + SQLite WAL (modernc.org/sqlite pure-Go driver) — busy timeouts, atomic claims, idempotency via Idempotency-Key
  • Ticker scheduler — polls every 15s, worker pool concurrency 2, threshold 200 gCO2e/kWh
  • Prometheus — /metrics (qharvest_task_queue_size, qharvest_carbon_intensity) + /healthz /readyz
  • Operator console at / — real-time grid chart + queue + WAL logs (see live demo)
  • Docker + Kubernetes manifests in deploy/ 
// pqc.go — integration boundary ready for ML-KEM
func EnvelopeIdentifier(data []byte) [32]byte {
    return sha256.Sum256(data) // placeholder for vetted ML-KEM/FIPS203
}
Enter fullscreen mode Exit fullscreen mode

How logging looks 

2026/08/15 12:01:45 Q-HARVEST listening on :8080
2026/08/15 12:02:00 scheduler: deferring queue at 270gCO2e/kWh
2026/08/15 12:02:15 scheduler: carbon intensity dropped to 145gCO2e/kWh
2026/08/15 12:02:15 scheduler: executing pqc-bulk-signing
Enter fullscreen mode Exit fullscreen mode

Why "harvest"?

Green computing is usually "use less." Quant Harvest is "use when clean." Same jobs, shifted to renewable windows — simplest decarbonization win for PQC migration.

Get involved

Apache-2.0. If you're Go/DevOps/crypto, try the live demo and roast the scheduler logic!

Top comments (0)