Store and Query
Industrial Operational History
Retain telemetry, production metrics, and equipment history in an industrial historian with explicit retention, aggregation, access, and deployment boundaries.
Why Proxus Data Storage?
General-purpose databases can become costly to operate for dense sensor history. Proxus uses a columnar time-series engine for industrial telemetry; capacity and query latency must be validated against the customer's tag density, retention, hardware, and workload. See storage architecture details.
- Workload-aware queries, index operational history by time and tag for the required analysis path.
- Columnar compression, reduce storage according to value types, cardinality, and retention design.
- Automatic retention, define TTL policies to purge old data without manual cleanup.
- Built-in aggregations, pre-compute averages, minimums, and maximums at ingest time.
What you can store
- Device tags and sensor readings (temperature, pressure, flow)
- Production metrics (OEE, cycle time, downtime events)
- Energy consumption (kWh, steam, water, gas)
- Quality measurements (dimensions, weights, defects)
- Equipment health (vibration, runtime hours, alarms)
See how storage fits into a larger governed layer for collection, context, routing, and enterprise delivery.
Storage Architecture
Data flows from edge gateways through the unified namespace into optimized storage partitions. Queries span hot and cold tiers seamlessly, recent data stays in memory while historical data lives on disk. Dive into ClickHouse docs for schema and retention setup.
Example retention policy
1-min aggregates: 1 year
1-hour aggregates: 5 years
Daily summaries: Forever
Balance storage cost and query speed by defining granularity tiers.
Query Performance
Dashboards, reports, and analytics tools can query the governed storage layer while downstream warehouses keep their own declared contracts and ownership. Review the FAQ for deployment guidance.
- Real-time + historical, query live UNS cache and long-term storage in one request.
- Tag-based filtering, select only the sensors and lines you need.
- SQL-compatible, use familiar aggregations and JOINs for complex analytics.
Use cases
- Trend analysis across production shifts
- Root cause investigation for downtime events
- Energy audits and ISO 50001 reporting
- Quality correlation studies (process vs. outcome)
- Predictive maintenance feature engineering
Integration & Compliance
Data storage integrates seamlessly with the rest of Proxus. Dashboards subscribe to live data while pulling historical context. Rule engines trigger on patterns across time windows. IT systems export archives for regulatory audits.
- Audit trails, track who queried what and when.
- Role-based access, scope queries by user, site, or line.
- Export formats, CSV, Parquet, or direct API access for BI tools.
Compliance inputs
- Retention windows and deletion policies
- Access roles and audit requirements
- Backup, restore, and evidence procedures
- Customer validation for applicable regulations
Technical and Commercial Evaluation
industrial historian and operational data storage evaluation guide
Operational problem it addresses
Operations teams need durable history with asset context, quality, retention, and access rules. A generic database alone does not define how plant signals become governed operational records.
Data sources it connects
- Timestamped PLC, OPC UA, Modbus, MQTT, and device measurements with quality metadata
- Production events, alarms, downtime states, counters, and contextual asset paths
- Selected calculated metrics and aggregates produced by edge or central workflows
How the data is processed
- 1.Validate and contextualize incoming values before writing operational records.
- 2.Partition and index data by time, tag, and the chosen asset model.
- 3.Apply retention, aggregation, and deletion policies defined for each data class.
- 4.Expose authorized historical queries or exports to operational and enterprise consumers.
Edge and outage behavior
Local collection and replay depend on the configured edge buffer, retention window, available disk, connector acknowledgement, and central storage health. High availability, backups, recovery objectives, and cross-site replication require deployment-specific design and testing.
Systems that consume the data
- Operations dashboards
- OEE and downtime analysis
- Energy reporting
- Quality investigation
- BI and data science
Security and deployment boundary
Storage can remain inside the customer-controlled environment. The customer owns infrastructure hardening, encryption and key policy, identity integration, backup destinations, retention approval, disaster recovery, and regulatory validation.
Technical validation and next step
When it is a good fit
- Operational history needs consistent asset context, quality, retention, and access policy.
- Teams need one governed source layer for dashboards and approved downstream exports.
- Capacity, recovery, and query targets can be benchmarked against a representative workload.
When it is not a good fit
- The requirement is only short-lived device buffering with no historical query workload.
- A certified records system is required without customer validation, procedures, and deployment evidence.
- Capacity or latency is being selected from an unqualified headline benchmark rather than a representative test.
Evaluation FAQ
How should historian capacity be sized?
Size it from tag count, value types, sampling and change rates, metadata, retention tiers, replication, query concurrency, and recovery objectives. A representative sustained workload test is required for a defensible commitment.
Does Proxus replace every enterprise data warehouse?
No. It provides governed operational history and context. Enterprise warehouses may still own cross-domain analytics, financial reporting, or long-term corporate data products.
Does retention configuration make a deployment compliant?
No. Retention is one technical control. Compliance also depends on validated procedures, identity and access, audit evidence, backup and restore, change control, and the applicable customer regulatory assessment.
FAQ
Common questions on storage capacity, query limits, and retention policies.
Capacity depends on tag density, sampling profile, value types, compression, retention tiers, replication, hardware, and query concurrency. It should be sized with a representative workload.
Write capacity is deployment-specific. Proxus validates sustained and peak behavior against representative hardware, tag density, polling or subscription profile, retention, and concurrent query load.
Yes. Define TTL (time-to-live) policies by table or partition to drop data after retention windows expire.
Yes. Export via REST API, ODBC, or direct database connectors for Power BI, Tableau, and Grafana.
Design your operational history around a real workload
From sensor telemetry to production analytics, store it once, query it fast, keep it compliant.