Enable better performance and reduce storage cost with InfluxDB 3.0
summary
InfluxDB 3.0 replaces the old TSM engine with a columnar design built on Rust, Apache Arrow, Parquet, DataFusion and Arrow Flight SQL, storing Parquet files in object storage and managing an in-memory hot layer for recent data. Claims include 4–6x disk footprint reduction, 45x write throughput on equal hardware and a quoted 90% reduction in storage cost. Queries on high-cardinality, recent data return in sub-second timeframes thanks to vectorized execution, column pruning and partition pruning.
AI-written summary · about
chapters (7)
- 0:00Introduction to InfluxDB
- 1:20Evolution of the platform
- 4:20Engineering InfluxDB 3.0
- 5:56Core technology components
- 8:25Performance and efficiency
- 11:50Benchmarking results
- 14:08Deployment and conclusion
related talks
42mColumnar Storage: Redefining Data Management for the Modern Era
13mNext-generation databases for next-generation workloads (PRT051)
40mInfluxDB storage engine internals
22mAt scale, everything is hard
26mColumnar Storage: Redefining Data Management for the Modern Era
12mInfluxDB for Industrial IoT
24mCracking the Cost Code: A Cost Optimized Approach to Distributed...
24mLeveraging the Apache Flight Python Client and InfluxDB