Case Study Deskcasestudydesk.com
Scientific ResearchSourced

Case Study: CERN cut SCADA archive storage 7× with TimescaleDB

CERN Case StudySourced & dated by Case Study Desk
Key facts · TL;DR
Company
CERN
Industry
Scientific Research
Challenge
Legacy Oracle-coupled SCADA archiver had technical debt and performance limits at scale
Headline result
CERN's NextGen Archiver on TimescaleDB delivered 7× storage reduction and 10-40× faster queries on compressed data

Key results

Storage reduction
78-95% space savings
10-40×
Faster queries on compressed data
77k
Rows/second write throughput
vs. 20k requirement

The challenge

CERN's legacy SCADA archiving system (the RDB Archiver, in use since 2008) was tightly coupled to Oracle, carried significant technical debt, and had rigid schemas and performance limits. It managed millions of daily time-series data points across more than 800 systems.

The solution

CERN deployed its NextGen Archiver (NGA) with a TimescaleDB backend, using the PostgreSQL extension's columnar compression, continuous aggregates and pluggable multi-database architecture.

For the read results using compression, we have observed pretty significant speed-ups, from 10x to 40x, depending on your query, range, and data frequency.

MZ
Martin Zemko
Software Engineer, CERN

The results, in context

CERN reported a 7× storage reduction (78-95% space savings depending on data type) and query speed-ups of 10-40× on compressed data depending on query, range and data frequency. Write throughput reached 77,000 rows per second against a 20,000 rows/second requirement. The system ran on roughly 500 systems at the time of writing, with full production planned for 2027.

Products used

Timescale TimescaleDB