The Future of Apache Hive and Hadoop 2.0: Bigger, Faster, Stronger
summary
Roadmap and hard numbers: Hive 0.10 ships decimals and rollups, 0.11 adds windowing and base optimizations, 0.12 targets richer types; Stinger initiative targets ~100x faster queries by adding a vectorized engine (early >30x per operator), a columnar ORC format, and smarter joins. Hadoop 2 brings Tez on YARN to avoid HDFS checkpoints and cut latency. Demo uses the Hortonworks sandbox with a JSON SerDe to expose arrays/maps and runs SQL (average battery by country) on semi‑structured data. Facebook runs Hive on 100+ PB and ~60k queries/day.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:25Apache Hive and SQL
- 7:45Hive vs traditional DBs
- 12:47Hive use cases
- 22:15Stinger initiative
- 31:48Hadoop 2.0 and YARN
- 38:16Hive 11 and 12 performance
- 42:53Demo: semi-structured data
- 53:25Q&A
related talks
15mThe Future of Hadoop: What Happened & What's Possible?
34mHive on Spark #bbuzz
15mThe Future of Data
02h 05mBig Data and the Data Lake from the Analyst Perspective
16mHadoop's Impact on the Future of Data Management
27mHive-Iceberg - Breaking the Ice: A Closer Look at Hive-Iceberg
35mProduction-Ready Flink and Hive Integration - what story you can tell now?
47mSolving Real & Big (Data) Problems Using Hadoop