Real-time analytics at cloud scale

The real-time analytics database built for the cloud. Sub-second SQL on streaming data with surprising efficiency.

Or start a free trial ->
Benchmarking Elasticsearch and Rockset: Rockset achieves up to 4X faster streaming data ingestionRead Now ->

Simple to start. Efficient to scale.

Peak usage of a real-world Rockset cluster.
Streaming Ingest
p50 Query Latency
Data Stored

Real-time speed

Serve low latency, high concurrency applications. Rockset’s Converged Index™, stored on RocksDB, gives you faster, more cost-effective data retrieval.

Search, aggregate and join any data at scale with millisecond latency SQL. Query data from CDC and event streams within 1-2 seconds of it being generated.

67% FASTER QUERIESCompare w/ Clickhouse->
Ritual logo
My engineering team is wowed by the query speed.

Kira Furuichi, Manager of Data Science and Analytics, Ritual

Cloud-native efficiency

The only real-time analytics database to independently scale compute-storage and compute-compute.

Avoid overprovisioning. Scale compute up and down on demand. Run multiple applications on shared, real-time data.

20-40% less computeCompare w/ Druid->
Meta logo
Rockset has proved to meet our requirements with its ability to scale, exist as an open and accessible cloud service, and query big datasets quickly.

Jane Xu, Software Engineer, Meta

Operational simplicity

Iterate faster with a fully managed database. No managing servers, clusters, shards and indexes.

Get started in minutes with built-in connectors for Kafka, DynamoDB, S3 and more. Run SQL on deeply nested JSON. Build apps in weeks, not months, regardless of the shape of your data.

20x faster developmentCompare w/ Elasticsearch->
Whatnot logo
Rockset delivered true real-time ingestion and queries, with sub-50 millisecond end-to-end latency that didn’t just match Elasticsearch, but did so at much lower operational effort and cost.

Emmanuel Fuentes, Head of Machine Learning and Data Platforms, Whatnot

Rockset Concepts, Design & Architecture ->

User-facing analytics

  • Real-time personalization icon
    Real-time personalization

    Harness behavioral data to create customized eCommerce experiences while a user’s session is still active.

  • Gaming icon

    Build real-time leaderboards, achieve instant in-game matchmaking, and implement dynamic pricing for in-game purchases.

  • Logistics icon

    Analyze thousands of data points in your fleet, the environment, and inventory management systems to take real-time actions.

  • Anomaly detection icon
    Anomaly detection

    Join streaming and historical data to identify fraud, sensor, and security anomalies as they happen.

  • IoT icon

    Use real-time sort, filter and search functionalities to analyze sensor data and trigger automations.

  • Vector search icon
    Vector search

    Fast KNN search with metadata filters to deliver relevant results, all in real time.

Ingest, transform, store and query in real time

With RocksetMillisecondsWithout RocksetMinutes or hours
With Rockset
Simple data flow diagram typical of a Rockset setup

Core features

Built in data connectors

Schemaless ingest using native connectors for Kafka, Kinesis, MSK, CDC streams, MongoDB, DynamoDB, S3, Postgres, MySQL, and more. Ingest millions of events per second.

Ingest transforms and rollups

Pre-aggregate and transform data at ingest time using SQL. Reduce the cost of storing and querying high-volume streaming data by up to 100x, and eliminate the need for separate real-time data pipelines.

In place updates

Efficiently insert, update, and delete data in place. Avoid the expensive and error prone workarounds inherent in append only databases.

High performance SQL JOINs

Avoid denormalization. Run fast, SQL JOINs on nested JSON, Parquet, Avro, and XML data formats. Rockset also supports streaming inner JOINs.

Converged Index™

Execute analytical queries on large datasets in milliseconds. All fields, including deeply nested fields, are automatically indexed in a Converged Index™, which includes an inverted index, columnar index and row index.

Compute compute separation

Independently scale compute and storage to avoid overprovisioning. Isolate compute for streaming ingest and queries. Run multiple apps on shared real time data.

Built on open source RocksDB

Leverage the same embedded storage engine trusted by Netflix, Linkedin, and Uber. RocksDB uses LSM trees to achieve the high volume write rates essential for ingesting streaming data.

Secure and reliable

Support for PII / PHI masking, role-based access control, SSO, IP allowlisting, MFA, audit logging, AWS PrivateLink, and more. Designed to be fault tolerant with cloud storage as the durability layer.

Ways to get started

Speak with an engineer

  • Speak with a senior database engineer with decades of experience
  • Receive a 1v1 consultation with best practices for cloud and streaming data architectures
  • Get a live demo tailored to your specific technical requirements
Get a demo->

Try it for yourself

  • $300 in credits, completely free, no credit card required
  • Access to sample data and hands-on-labs
  • Actively monitored Rockset community for support and feature requests
Start a free trial->