Skip to content

CommP

CommP (Piece Commitment) configuration.

Key Default Env Dynamic
pdp.aggregation.commp.job_queue.workers runtime.NumCPU() PIRI_PDP_AGGREGATION_COMMP_JOB_QUEUE_WORKERS No
pdp.aggregation.commp.job_queue.retries 50 PIRI_PDP_AGGREGATION_COMMP_JOB_QUEUE_RETRIES No
pdp.aggregation.commp.job_queue.retry_delay 10s PIRI_PDP_AGGREGATION_COMMP_JOB_QUEUE_RETRY_DELAY No

Overview

CommP calculation is the first stage of the aggregation pipeline. When a blob is uploaded, it enters this queue to have its Piece Commitment hash calculated. This hash is required for creating aggregates that can be submitted to the blockchain.

Performance Note: CommP calculation is CPU-intensive. Each worker consumes significant CPU resources while processing a blob.

Fields

job_queue.workers

Number of concurrent CommP calculations. Defaults to the number of CPU cores.

  • Higher values: Faster throughput when many blobs arrive simultaneously, but higher CPU usage
  • Lower values: Reduced CPU load, but blobs queue longer during high ingest periods

job_queue.retries

Maximum retry attempts before a blob is moved to the dead-letter queue.

job_queue.retry_delay

Wait time between retry attempts after a failure.

TOML

[pdp.aggregation.commp.job_queue]
workers = 4        # Limit to 4 cores for shared environments
retries = 50
retry_delay = "10s"