Technical White Paper
PlantPulse Industrial DataLake Technical White Paper. This document is intended for technical reviewers (data architects, infrastructure/DB operators, and security reviewers) and explains, through structure, metrics, configuration, and field measurements, how an industrial data lake differs from a general-purpose lakehouse.
Download
| Language | Pages | File |
|---|---|---|
| Korean | 22 pages | PlantPulse Industrial DataLake Technical White Paper (PDF) |
| English | 22 pages | PlantPulse Industrial DataLake Technical White Paper (PDF) |
Revision 2026.09.
What's Inside
| Chapter | Contents |
|---|---|
| 01–02 | The three barriers (Historian · General-Purpose Lakehouse · DIY Integration) and five design principles |
| 03–04 | Batch · Ports · Startup sequence · Four image sources and verification set |
| 05 | Data lifecycle — Hot (in-memory) · Warm (Aster TSDB ring) · Cold (Iceberg) |
| 06 | Aster TSDB — Proprietary time series database. A KOPENS engine layered on Cassandra 6.0 with compression, retention, and time series computation — 7.1× storage reduction, 21 CQL time series functions, transparent tiered reads |
| 07–09 | Schema · TSE · Time series DB comparison |
| 10–11 | Lakehouse · Messaging (Kafka · Aster MQTT · Two buffers) — startup denied without authenticator · Per-drop reason counters |
| 12 | Aster CEP — Proprietary real-time decision engine. EQL → Common IR → In-server engine, SQL:2016 MATCH_RECOGNIZE engine patch vs. Flink, upstream equivalence CI gate |
| 13–14 | Query layer — Data Gateway. Four languages · Endpoints · Parameter binding, and control · Error contract · Auditing · JDBC driver |
| 15–16 | Scaling and replication · Operations (Console · pd · JSON contract) |
| 17–18 | Protection (Backup · Recovery · Upgrade) · Measurements from running instances |
| 19–20 | Comparison with general-purpose data lakes · Sizing · Sources |
How to Read This Document
This is not sales material. All metrics are taken solely from repository configuration, schema, benchmark documents, and measurements from running instances, with every source cited on page 22. The storage engine benchmark (7.1× compression · 233k rows/s) is the publicly measured value from kopens/aster-tsdb.
The default out-of-box retention period for raw tag values is 3650 days (10 years). As this is the largest storage axis, it is often shortened in the field — for details on values and how to change them, see Data Retention Period. For capacity planning, refer to Data Capacity Planning as well.
See Also
Download the Proposal · Comparison with Competitive Products
To delve deeper into chapters 12–14 of this White Paper, refer to the CEP and Data Gateway sections.