Every decision starts with reliable data.

The Data Layer is Layer 2 of the 5-layer Rayven Platform stack. It sits directly above the Integration Layer and is responsible for storing, transforming, contextualising, and governing all operational data that enters the platform. It provides the unified, queryable data foundation that every layer above it — analytics, AI, automation, and dashboards — depends on.

Rayven uses a native time-series data store optimised for high-frequency sensor and operational data. It supports efficient compression of repetitive industrial readings, configurable downsampling for different retention tiers, and fast range queries over millions of data points. Time-series data is stored alongside event, transactional, and document-style data in a unified model.

The Data Layer includes a full transformation engine that can clean, normalise, aggregate, filter, join, and enrich data either at ingestion (streaming transforms) or on-demand (batch). Transformation rules are configured in the platform without custom code and can be versioned, audited, and reused across multiple data flows.

When data arrives from different source systems, the Integration Layer routes it into the Data Layer where a unified schema maps each field to a common data model. Asset identifiers, timestamps, units of measure, and business hierarchy fields are standardised across all sources — so a temperature reading from a SCADA system and one from an IoT device are stored and queried identically.

Data quality rules can be applied at ingestion or as standing quality checks on stored data. This includes range validation (flagging values outside expected bounds), completeness checks (detecting missing fields or gaps in time-series), deduplication, and statistical anomaly detection. Records that fail quality checks can be quarantined, flagged, or routed to a review queue rather than silently dropped.

Contextualisation is the process of enriching raw operational data with the business meaning that makes it interpretable. A raw sensor value becomes useful when it is tagged with the asset it measures, the site it belongs to, the engineering unit it represents, and the business process it relates to. Rayven applies contextualisation rules at the Data Layer so every consumer — dashboard, workflow, AI model — receives pre-enriched data without needing to rediscover context each time.

Yes. Rayven includes master data management (MDM) capabilities that allow you to define and maintain reference data sets — assets, locations, product codes, personnel, equipment hierarchies, and more. MDM records can be used to automatically enrich incoming operational data and to drive consistent naming and classification across reports, dashboards, and workflows.

The Integration Layer (Layer 1) handles ingestion — collecting data from external sources and routing it into the platform. The Data Layer (Layer 2) handles everything that happens to data once it is inside: storage, transformation, contextualisation, quality management, and governance. The Integration Layer feeds the Data Layer; the Data Layer feeds everything above it. They are designed as complementary layers with clean separation of concerns.

Yes. The Data Layer exposes a queryable interface accessible to all Rayven modules and to authorised external tools via the API Layer. You can run ad-hoc queries, schedule data exports to file or external systems, and connect third-party BI tools directly. All queries are governed by the same access control rules as the rest of the platform.

Rayven supports configurable data retention policies, including hot (full resolution), warm (downsampled), and cold (archive) tiers with automatic migration between them. Retention periods are configured per data type and business requirement. Industrial-scale retention is efficient by design — time-series compression and tiered storage mean you can retain years of sensor data without the cost of storing it all at full resolution.