Skip to main content

Analytics Architecture

TrackVision's analytics layer is built on an HTAP (Hybrid Transactional and Analytical Processing) database — the same database that stores your operational traceability data also powers your analytical queries. This means there is no ETL pipeline, no data warehouse to sync, and no lag between an event being recorded and it appearing in your analytics. You query live data.

What HTAP Means in Practice​

Traditional architectures separate transactional databases (fast writes, low latency) from analytical databases (columnar storage, optimised for aggregation). Data moves between them via batch ETL jobs, introducing hours or days of lag. HTAP eliminates this split by maintaining both row-oriented and columnar representations of the same data simultaneously.

For supply chain analytics this matters because:

  • A shipment scanned at a distribution centre is immediately queryable in a dashboard — no overnight batch run
  • You can run complex aggregations (e.g. which lots touched a recalled supplier across 6 months of EPCIS events) without contending with live transactional workloads
  • There is a single source of truth; your analytics numbers match your operational data exactly

The database scales horizontally to handle petabyte-scale datasets, which is relevant if you are ingesting high-frequency serialised scan events across a large product portfolio.

Infrastructure Properties​

Elastic horizontal scaling — the database scales across nodes transparently, without manual sharding or schema changes. Storage and compute can be scaled independently.

ACID transactions at scale — full transactional consistency is maintained even under high write volumes, which is important when ingesting bulk EPCIS events while simultaneously running analytical queries.

Cross-zone replication — data is replicated across multiple availability zones automatically, providing resilience against infrastructure failure without additional configuration.

Encryption in transit and at rest — all data is encrypted. Network isolation is enforced at the infrastructure level.

No ETL, no stale data — because transactional and analytical processing share the same storage layer, there is no synchronisation job that can fall behind or fail silently. A query run at any moment reflects the current state of the data.

Why This Matters for Traceability​

Supply chain traceability generates high write volumes (scan events, EPCIS captures, sensor readings) and also demands complex analytical queries that span large time ranges and join across many entities. These two workloads typically pull in opposite directions — transactional systems prioritise low-latency writes, analytical systems prioritise throughput on large scans.

HTAP handles both without requiring the operator to maintain separate systems, synchronisation pipelines, or tolerate stale analytics. For recall investigations, regulatory audits, or real-time operational monitoring, this means results reflect reality at the time of the query.