CoursesChallenges

Real-Time OLAP

User-facing analytics with sub-second p99 on fresh data. Druid and Pinot side by side: their role-based architectures, segments as the unit of storage, streaming and batch ingestion, ingest-time rollup, Druid bitmap indexes, Pinot inverted/sorted/range indexes and the star-tree, Pinot upserts, tiered and deep storage, the two query surfaces, and a decision framework grounded in tenancy, upserts, joins, and ops cost. Ends with a real-time dashboard backend capstone.

Advanced15 chapters· 4h 18m· in Query Engines & OLAP

Course content

  1. 01The Real-Time OLAP WorkloadFree
  2. 02Druid Architecture🔒
  3. 03Pinot Architecture🔒
  4. 04Segments - The Unit of Storage🔒
  5. 05Ingestion in Druid🔒
  6. 06Ingestion in Pinot🔒
  7. 07Rollup & Pre-Aggregation at Ingest🔒
  8. 08Indexes - Druid Bitmaps🔒
  9. 09Indexes - Pinot Inverted, Sorted, Range🔒
  10. 10Star-Tree Index (Pinot)🔒
  11. 11Upserts in Pinot🔒
  12. 12Tiered Storage & Deep Storage🔒
  13. 13Querying - Druid SQL/Native vs Pinot SQL🔒
  14. 14Druid vs Pinot Decision Framework🔒
  15. 15Capstone - Real-Time Dashboard Backend🔒

Prerequisites

↗Query Engine Foundations

Read the first chapter free

Start reading now — no account required for the free chapters.

Start: The Real-Time OLAP Workload →More in Query Engines & OLAP