LoginStart for freeBook a demo
Multi-Stage Query Engine

Analyze your data at your speed

Run complex SQL queries without sacrificing blazing-fast query performance, gaining the business insights your teams need in real-time.

Book a demo
Start for free
Multi Stage Query Engine

Deliver fast and accurate insights across large datasets

Fastest Performance Yellow

Improve query performance

64x64 Cost Icon Purple V1

Allocate resources more effectively without increasing costs

64x64 Scalability Icon Blue V1

Scale to handle increased data volumes and query complexity

Run complex queries with native query-time joins

Unlock timely and accurate insights from your data in motion and power your real-time analytics application.

Joins Support

Optimize query performance and the execution of joins at scale with native support for serving joins with subseconds latency. The Multi-Stage Query Engine supports all three types of join strategies: Broadcast Joins, Shuffle Hash Distributed Joins, and Lookup Joins. With this capability, StarTree covers the entire spectrum of data joins, providing full coverage from user-facing analytics all the way up to ad hoc analytics.

Query Feature Joins V1

Multi-Stage Execution Model

The multi-stage execution model is designed to handle complex multi-stage data processing. It includes an intermediate compute stage (consisting of a set of processing servers and a data exchange mechanism) that allows StarTree to handle more complex processing requirements by offloading computation from the brokers.

Multi Stage Query Engine

Indexing and Pruning

StarTree offers a rich set of indexing and pruning techniques to speed up query processing on individual tables. These techniques help reduce the overhead of scanning and aggregations to improve query latency and throughput.

Query Feature Indexing V1

Data Layout

Improve join performance by adapting based on the data layout. StarTree optimizes joins to account for data that is partitioned but not co-located, data that is partitioned and co-located, and random layouts (data is neither partitioned nor co-located).

Query Feature Layout V1
Stripe Logo White

“We migrated our old time series aggregation system from a different framework to Apache Pinot. With Pinot, we can ingest an entire stream of data and send multiple queries at it, and it will be equally performant. This offers a lot of flexibility — it allows us to lower the cost of operating, and opens up real-time analytics to folks that weren’t Scala experts but know how to write SQL queries.”

Johan Adami Stripe
Johan Adami
Software Engineer
keep exploring

The most powerful real-time analytics platform

Simplified tiered storage

Analyze all your data, new and historical, without slowing down queries or blowing your budget. StarTree automatically stores the data you access most often for super-fast results.
Tiered Storage Horizontal V1

Powerful, flexible indexing

Harness extensive indexing to analyze your data, any way you want. Unlike other techniques which work on a single column, the star-tree index is built on multiple columns, using pre-aggregated results.
Simplified Tiered Storage 1

Scalable real-time upserts

Upserts are two operations in one. If a data record doesn’t exist, it performs an insert. If it does already exist, it performs an update. StarTree can perform upserts scalably at millons of rows per second.
Scalable Real Time Upserts

Autoscaling minions

Save costs on your cloud infrastructure with autoscaling minions. Ensure tasks are processed efficiently and drop the legwork of repartitioning data post-ingestion and merging smaller segments into larger ones.
Auto Scaling Minions 2

Ready to deploy real-time analytics?

Start for free or book a demo with our team.

Start for free
Book a demo