PushdownDB: Accelerating a DBMS using S3 Computation
Preprint
Abstract
This paper studies the effectiveness of pushing parts of DBMS analytics queries into the Simple Storage Service (S3) engine of Amazon Web Services (AWS), using a recently released capability called S3 Select. We show that some DBMS primitives (filter, projection, aggregation) can always be cost-effectively moved into S3. Other more complex operations (join, top-K, group-by) require reimplementation to take advantage of S3 Select and are often candidates for pushdown. We demonstrate these capabilities through experimentation using a new DBMS that we developed, PushdownDB. Experimentation with a collection of queries including TPC-H queries shows that PushdownDB is on average 30% cheaper and 6.7X faster than a baseline that does not use S3 Select.Keywords
All Related Versions
This publication has 12 references indexed in Scilit:
- Biscuit: A Framework for Near-Data Processing of Big Data Workloads2016
- The Snowflake Elastic Data Warehouse2016
- HRL: Efficient and flexible reconfigurable logic for near-data processing2016
- Amazon Redshift and the Case for Simpler Data Warehouses2015
- Near-Data Processing: Insights from a MICRO-46 WorkshopIEEE Micro, 2014
- Hive - a petabyte scale data warehouse using Hadoop2010
- Scalable Bloom FiltersInformation Processing Letters, 2007
- Active disks for large-scale data processingComputer, 2001
- A case for intelligent disks (IDISKs)ACM SIGMOD Record, 1998
- Universal classes of hash functionsJournal of Computer and System Sciences, 1979