Lightning Talk - Parquet: The Pillar of Modern Telemetry...
summary
Parquet is a binary columnar storage format that groups a few hundred thousand rows into row groups, stores each field physically together as Parquet columnar format, and uses dictionary encoding to shrink repeated values. The claim is concrete: real-world telemetry (logs, traces, metrics) in the uploader’s use cases compresses by about 90% and retains queryable structure, so high-cardinality metric fields become efficiently stored columns rather than exploding storage.
AI-written summary · about
related talks
39mHow to avoid columnar calamities: what no one told you about Apache Parquet
46mEfficient Data formats for Analytics with Parquet and Arrow
26mColumnar Storage: Redefining Data Management for the Modern Era
29mBeyond TSDB: Unlocking Prometheus with Parquet for Modern Sc...
37mBenefits of columnar data layout
47mAnalytics at the speed of Direct Lake
26mPetabyte Scale Datalake Table Management with Ray, Arrow, Parquet, and S3
31mMapping FHIR data to parquet for analytics and machine learning