Best Columnar Databases for Amazon S3

More customers pick Amazon Redshift than any other cloud data warehouse. Redshift powers analytical workloads for Fortune 500 companies, startups, and everything in between. Companies like Lyft have grown with Redshift from startups to multi-billion dollar enterprises. No other data warehouse makes it as easy to gain new insights from all your data. With Redshift you can query petabytes of structured and semi-structured data across your data warehouse, operational database, and your data lake using standard SQL. Redshift lets you easily save the results of your queries back to your S3 data lake using open formats like Apache Parquet to further analyze from other analytics services like Amazon EMR, Amazon Athena, and Amazon SageMaker. Redshift is the world’s fastest cloud data warehouse and gets faster every year. For performance intensive workloads you can use the new RA3 instances to get up to 3x the performance of any cloud data warehouse.

StarTree, powered by Apache Pinot™, is a fully managed real-time analytics platform built for customer-facing applications that demand instant insights on the freshest data. Unlike traditional data warehouses or OLTP databases—optimized for back-office reporting or transactions—StarTree is engineered for real-time OLAP at true scale, meaning: - Data Volume: query performance sustained at petabyte scale - Ingest Rates: millions of events per second, continuously indexed for freshness - Concurrency: thousands to millions of simultaneous users served with sub-second latency With StarTree, businesses deliver always-fresh insights at interactive speed, enabling applications that personalize, monitor, and act in real time.

Pinot is designed to answer OLAP queries with low latency on immutable data. Pluggable indexing technologies - Sorted Index, Bitmap Index, Inverted Index. Joins are currently not supported, but this problem can be overcome by using Trino or PrestoDB for querying. SQL like language that supports selection, aggregation, filtering, group by, order by, distinct queries on data. Consist of of both offline and real-time table. Use real-time table only to cover segments for which offline data may not be available yet. Detect the right anomalies by customizing anomaly detect flow and notification flow.