Reinventing Amazon Redshift
summary
Covers how Amazon Redshift has been redesigned to improve performance and scalability by changing its core architecture and execution engine. Describes specific technologies and techniques used in the redesign (new storage, query processing, and resource management components) and explains how those components work together to address bottlenecks in the original Redshift system.
AI-written summary · about
chapters (22)
- 0:00Intro
- 2:36Challenges of data analytics at scale
- 3:12The Dark Data Problem
- 4:04Amazon Redshift-most popular & fastest cloud data warehouse
- 5:19The Amazon Redshift architecture - 2021
- 7:32Executing a query in Amazon Redshift
- 10:48Code generation, Prefetching, Vectorization
- 14:12Continuous telemetry and benchmarking
- 15:27Compilation-as-a-Service
- 20:58Linear scalability to the P8 scale
- 22:14Diverging speeds in hardware
- 22:49AQUA - A push-down & scale-out computation layer
- 24:23Redshift Managed Storage
- 24:53Cross-AZ cluster recovery
- 28:26Compute elasticity improvements
- 34:08Detecting Errors
- 34:29Detecting Gray Errors
- 37:32Automatic Table Optimization
- 39:04Automated workload management
- 46:13Ingesting and Querying Semistructured Data
- 47:44Amazon Redshift Spectrum
- 49:00The evolution of Amazon Redshift
