Skip to main content
StreamNative Kafka Service is a fully managed, native Apache Kafka service built on the Lakestream architecture and powered by the Ursa Engine. It is not a Kafka compatibility layer — it runs native Apache Kafka with a lakehouse-native storage engine that delivers leaderless, compute-storage separated architecture with up to 95% lower infrastructure costs. Your existing Kafka clients, Kafka Connect connectors, and Kafka Streams applications work without any code changes. Just point them at your StreamNative Kafka endpoint.

Key advantages

Native Apache Kafka

Runs native Apache Kafka — not a compatibility layer. Use any standard Kafka client, tool, or ecosystem without modification.

Up to 95% Lower Cost

Leaderless architecture with compute/storage separation eliminates cross-AZ replication fees. Data writes directly to object storage.

Lakehouse-Native

Every topic is simultaneously an Iceberg or Delta Lake table. Zero-copy streaming into your lakehouse with no connectors or ETL.

Zero Operations

Fully managed. No brokers to provision, no ZooKeeper to manage, no partitions to rebalance. Auto-scaling handles the rest.

Coming from an existing Kafka deployment?

From Amazon MSK

Lower cost with lakehouse-native storage. No cross-AZ replication fees. Same Kafka APIs.

From Confluent

Open formats, no vendor lock-in. Data stored in Iceberg and Delta Lake on your object storage.

From Self-Managed

Zero operational overhead. Auto-scaling, managed infrastructure, and enterprise-grade security.

Deployment options

StreamNative Kafka Service is available in multiple deployment configurations:
OptionDescriptionBest for
ServerlessFully managed, pay-per-useDevelopment, testing, small workloads
DedicatedDedicated infrastructure, predictable performanceProduction workloads
BYOCRuns in your cloud account, managed by StreamNativeData sovereignty, compliance
BYOC ProAdvanced BYOC with custom networking and DNSEnterprise, regulated industries
See Cluster Types for detailed feature comparison.

How it works

StreamNative Kafka Service runs on the Ursa Engine, a cloud-native storage engine at the heart of the Lakestream architecture.
  • Stateless brokers: Kafka brokers are stateless and leaderless. Any broker can handle produce or fetch requests for any partition.
  • Object storage: Data writes directly to S3, GCS, or Azure Blob Storage. No local disks, no inter-broker replication.
  • Lakehouse-native: Data is stored in open table formats (Iceberg, Delta Lake) on your object storage, queryable by any analytics engine.
Learn more about the Lakestream architecture and the Ursa Engine.

Get started

Quick Start

Create a Kafka cluster and produce your first message in under 5 minutes.

Migration Guide

Migrate from MSK, Confluent, or self-managed Kafka with zero code changes.

Kafka Compatibility

Check supported Kafka APIs, client versions, and feature compatibility.

Choose Kafka or Pulsar

Not sure which protocol fits your workload? Compare Kafka and Pulsar side by side.