How innovation happens on Amazon Redshift
summary
Amazon Redshift is a Postgres‑based MPP data warehouse reworked with columnar storage and per‑column compression, in‑memory zone maps and flexible distribution styles (including a FULL copy per node) to avoid IO and data movement. Redshift Spectrum lets the cluster query S3 directly; one demo scans 20 years (~140 TB/yr, 190M files, 15k partitions) and reduces work by ~3.5 billion× to return results in ~3 minutes. Pricing is pay‑per‑scan (~$5 per TB), the service iterates every two weeks, and over 120 significant features land since launch.
AI-written summary · about
chapters (13)
- 0:00Start
- 3:14Press Release
- 7:56Structured Data Processing
- 10:17Redshift Cluster Architecture
- 17:28Data Distribution Distribution style is a table property which dictates how that table's data is distributed throughout the cluster
- 20:24The Dark Data Problem Most generated data is unavailable for analysis
- 20:49The tyranny of "OR"
- 22:40Life of a query
- 28:29Amazon Redshift Spectrum is cost-effective
- 29:23Amazon Redshift Spectrum uses standard SQL
- 29:40Defining External Schema and Creating Tables Define an external schema in Amazon Redshift using the Amazon Athena data catalog or your own Apache Hive Metastore
- 31:19Converting to Parquet and ORC using Amazon EMR
- 35:50The Emerging Analytics Architecture
related talks
26mAmazon Redshift Reinvented
48mData warehousing reinvented for today’s needs
23mOne day with Redshift
41mOwning Your Own (Data) Lake House
22mReference architectures for Amazon Redshift data warehouses
43mDeep dive into Amazon Aurora and its innovations
22mRedshift Serverless - easily run big data analytics at scale
19mPostgreSQL at Amazon