Don't cry over spilled records: Memory elasticity of data-parallel applications and its application to cluster scheduling.
Calin IorgulescuFlorin DinuAunn RazaWajih Ul HassanWilly ZwaenepoelPublished in: USENIX Annual Technical Conference (2017)
Keyphrases
- data points
- database
- data sets
- data objects
- synthetic data
- data transfer
- databases
- raw data
- original data
- data analysis
- high quality
- training data
- computing power
- computational power
- data records
- parallel processing
- memory footprint
- attribute values
- computer systems
- data collection
- data processing
- image data
- knowledge discovery