TVStore: Automatically Bounding Time Series Storage via Time-Varying Compression
summary
TVStore applies a new technique called time-varying compression to bound storage for extreme-scale time series: instead of pointwise compression it fits time-parameterized functions to older data and triggers compression according to provable rules that guarantee a storage cap. The system is implemented by modifying an open-source time-series database and is evaluated on synthetic and real-world traces for storage boundedness, write/query efficiency, and the evolving compression patterns as data ages.
AI-written summary · about
chapters (6)
- 0:00Intro
- 0:25Time Series Management: Popularity & Volume
- 1:10Motivation 1: Limited Resources and Expenses
- 5:31Time-varying Compression
- 8:30Design choices for automatic bounding
- 11:25Experimental Settings
related talks
50mTime Series with SQL Server 2022
15mAn Efficient Cloud Storage Model with Compacted Metadata Management for Performance...
24mTSDB Data Compactions on Scale Without Headache: Now & Future
39mA New Data Model for Time Series Data Processing
30mBuilding a Time-Series Database for Point-in-Time Lookups
42mAccessible Time Series data
44mIntroduction to Time Series
29mCatching Waves with Time-Series Data