On the path to data Nirvana: Supporting OLTP and OLAP on Hadoop
summary
Claims a single, scale‑out RDBMS called Splice Machine runs OLTP and OLAP side‑by‑side by using HBase for OLTP and Spark for OLAP, with a cost‑based optimizer that routes queries and uses cgroups plus Spark fair scheduling for resource isolation. Built from Apache Derby parsing, HBase storage and Spark execution, it reports 10x price‑performance versus Oracle on a Unica workload, supports distributed ACID MVCC (large transactions, millions of updates), and connects to Kafka and federated data sources.
AI-written summary · about
chapters (10)
- 0:00Introduction to mixed workload databases
- 2:20Limitations of traditional architectures
- 4:50Introducing Lambda and relational Lambda
- 7:42Use cases for Splice Machine
- 13:03Dual engine architecture explained
- 16:01How Splice Machine is built
- 23:13Query management and performance tuning
- 24:57Federated query support
- 27:32ACID transactions on Hadoop
- 31:23Summary and Q&A
related talks
31mUnified MLOps: Feature Stores & Model Deployment
28mBreakout Session
50mDesigning a scalable data platform based on Apache Spark/Iceberg
45mFrom Big Data to Big Information
05mClose Encounters with the Third Kind of Database
07mindustries needing data disruption
38mKeynote | Data Warehouse, Data Lakes, What's Next?
39mAccelerate big data analytics with Apache Kylin