Database In-Memory: Real-time analytics, supercharged
summary
Oracle Database In-Memory keeps the row store and adds a pure columnar SGA area you enable with inmemory_size and alter table ... inmemory, letting the optimizer choose column or row access. Use the in-memory advisor and Compression Advisor to target objects; customers report up to 100x faster reports, 33x–41x compression (some tables), 2x average runtime drop on Exadata, and thousands of hours saved. Features include SIMD vector scans, in-memory storage indexes, in-memory duplication on Exadata, double-buffering and incremental repopulation for HTAP workloads.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:24Why use Database InMemory
- 2:22What is Database InMemory
- 5:07How does Database InMemory work
- 10:08Why not cache the table in the buffer cache
- 11:04How do you know which queries will benefit
- 12:03Does it work with other database features
- 14:58Whats new with database
- 15:53Realtime enterprises
- 16:26Hardware trends
- 16:49Three customers
- 17:25Lufthansa
- 18:31How did InMemory help
- 20:20Hisense
- 21:11Benefits of InMemory
- 23:17Staff
related talks
01h 34mWhat's Special About In-Memory Database Systems
22mOracle Exadata: Architecture and internals technical deep dive
19mSpeed up transparent performance with Exadata: What, when, how, and why
59mIndex Internals: RowStore, ColumnStore, and NoStore!
32mJava's Ignored Potential: Supercharge Database Apps with Java In-Memory Process
15mAccelerate Application Response Times with True Cache
51mIn-Memory Computing - The Big Picture
29mOracle TimesTen: extreme performance for latency-critical applications