Process Historian Software | Products

Overview

Process Historian is a purpose-built time-series database for industrial automation — engineered to record every tag on every PLC, every sensor on every line, every 100 milliseconds, for years, without falling over. Where a generic SQL database chokes after a few million rows, the historian holds billions of compressed samples and replays any 24-hour window in under a second.

Designed for plants that already have a SCADA / DCS but need long-term forensic history: batch comparison, golden-batch analysis, regulatory reporting (FDA, ISA-88), ML feature engineering, and root-cause investigation after an alarm or yield deviation.

Billions of samples — 2,800+ tags @ 100 ms scan = 24 billion samples / year
15–30:1 lossless compression — swinging-door + delta + Gorilla algorithms
Hot / warm / cold tiers — NVMe → SSD → object store (S3 / MinIO)
SQL + REST + Python + OPC HDA — query from any tool your engineers already use

How It Works

A lightweight collector pulls raw values from PLCs and SCADA via OPC UA / OPC HDA / MQTT / REST. The compression engine applies swinging-door, delta-of-delta, and dictionary encoding before writing to the hot tier. Continuous roll-ups generate 1-minute, 5-minute, 1-hour, and 1-day aggregates so dashboards never scan raw data.

Collection — OPC UA / HDA, MQTT, Modbus, REST, Sparkplug B at 10 ms–60 s rate
Compression — swinging-door (engineering deadband) + Gorilla (binary) + dictionary tags
Tiering — hot 30 d raw, warm 1 y roll-ups, cold 7+ y on object store
Query & replay — cursor scrub through any historical window at up to 100× real time
Federate — query across multiple plants or sites as a single virtual tag space

Features

Multi-trend chart — overlay any number of tags, dual y-axis, statistical bands
Replay scrubber — rewind / play / fast-forward at 1× / 4× / 100× speed
Batch & campaign view — ISA-88 batch IDs, golden-batch overlay, deviation highlighting
Asset model — logical tree (Site → Area → Line → Asset → Tag) instead of flat names
Alarm history — co-located alarm log with sequence-of-events at 1 ms resolution
Python / Jupyter SDK — pandas DataFrames in 2 lines, native ML integration
OPC HDA + REST — standards-compliant for any third-party SCADA / MES / BI tool
Audit & CFR 11 — immutable journal, electronic signatures, 21 CFR Part 11 compliant

Use Cases

Golden-batch analysis — pharma / food: compare every batch against the proven gold standard
Regulatory reporting — FDA / EPA / ISO 50001 reports auto-generated from history
Root-cause investigation — zoom into the 100 ms before a trip; see every related tag
Energy / sustainability — kWh per ton, GHG accounting, ISO 50001 baseline
Predictive maintenance — train ML models on years of vibration / current trends
OEE / KPI reporting — daily, weekly, monthly OEE / yield / downtime PDFs by email

Specifications

Ingest: 500,000 values/s per node; clustered to millions; 10 ms minimum sample rate
Tags: 1 million tags per server; 100 million across a federated cluster
Compression: 15–30:1 typical; swinging-door + Gorilla + ZSTD; lossless option
Query: < 100 ms for 24-h trend of 4 tags; SQL, REST, gRPC, OPC HDA, Python
Retention: hot NVMe 30 d, warm SSD 1 y, cold S3/MinIO 7+ y
HA: active-active cluster, RPO 0, RTO < 5 s, geographic replication
Connectors: OPC UA, OPC HDA, MQTT, Sparkplug B, Modbus, Profinet, REST, CSV import
Compliance: 21 CFR Part 11, ISA-88, ISO 50001, GAMP 5 documentation