Striim

Overview

Striim is a real-time data integration and streaming platform, and the difference from a pure replication tool is that it processes data in flight as well as moving it.

Capture is non-intrusive and log-based, with a named reader per source: Oracle Reader and OJet, MS SQL Reader and MSJet, MySQL, MariaDB, PostgreSQL, MongoDB, HP NonStop and Salesforce, all recoverable.

Unusually it also publishes query-based alternatives, a JDBC Database Reader and an Incremental Batch Reader, which matters for sources with no usable log. Debezium has no equivalent path.

The platform part is streaming SQL. You create an application with SQL, or through wizards and a flow designer, and get windowing, enrichment, complex event processing and PII obfuscation in the pipeline rather than afterwards.

It comes in three editions: a fully managed Striim Cloud on AWS, Azure or Google Cloud, a self-managed Striim Platform, and a free serverless Developer sandbox.

One sizing caveat matters more than any feature. Multi-server deployment, automatic failover and load rebalancing are available on Striim Platform and Cloud Mission Critical, and explicitly not on Cloud Enterprise, which is single-server.

Its delivery guarantees are the most clearly published of the commercial options. At-least-once is the default, with exactly-once named per writer for Kafka, Kinesis, JDBC and Cassandra, using recovery checkpoints every ten seconds.

Schema evolution is a named feature with candid per-target limits: Snowflake supports only adding and dropping columns, and creating a schema works only on the Iceberg writer.

Monitoring is a strength, with end-to-end lag sampled every five seconds and retained a day, and alerts to email, Slack, Teams or the web console.

The heritage is worth knowing in this category: Striim was founded by the core team from GoldenGate Software, which Oracle acquired in 2009, so it shares founding DNA with Oracle GoldenGate.

Key features and capabilities

The same headings are used for every change data capture & replication entry, so two tools can be read side by side.

How changes are captured

• Log-based capture through named readers, all recoverable and supporting replication
• Oracle through Oracle Reader or OJet, SQL Server through MS SQL Reader or MSJet
• Query-based alternatives are first-class, a JDBC Database Reader and an Incremental Batch Reader
• Wizards generate an initial-load application alongside the change-capture one
• A schema conversion utility ships with Striim Platform but not with Striim Cloud

Source databases

• Change capture for Oracle, SQL Server, MySQL, MariaDB, PostgreSQL, MongoDB, HP NonStop and Salesforce
• File and object sources on S3, ADLS, Google Cloud Storage and HDFS, in Avro, JSON, Parquet and XML
• API sources including Salesforce, Stripe and Zendesk
• No mainframe reader and no Db2 reader appear in the readers overview
• Supported versions are not published on the readers overview, unlike Qlik's explicit matrix

Targets

• Warehouses including Snowflake, BigQuery, Databricks, Redshift, Azure Synapse and Fabric
• Streaming to Kafka, Kinesis, Event Hubs, Google Pub/Sub, JMS and MQTT
• Object storage on S3, Google Cloud Storage, ADLS, Azure Blob and Fabric Lakehouse
• Databases and NoSQL including JDBC, Cassandra, MongoDB, HBase and Cosmos DB
• An Iceberg writer exists in the schema documentation but not in the writers overview; Delta and Hudi are not published

Delivery and consistency guarantees

• At-least-once by default, with exactly-once named per writer rather than claimed platform-wide
• Exactly-once writers are Kafka, Kinesis, JDBC and Cassandra; MongoDB and Event Hubs can be either
• Recovery checkpoints are written every ten seconds once the previous one completes
• The Kafka writer combines Kafka transactions with Striim checkpointing, needing an extra checkpoint topic
• Ordering and transaction-boundary preservation are not published as explicit guarantees

Handling schema changes

• Schema evolution is a named feature, keeping the target in sync without interrupting the application
• Supported DDL covers create, alter including add, modify, drop and keys, and drop table
• Per-target limits are published candidly; Snowflake supports only adding and dropping columns
• Creating a schema is supported only by the Iceberg writer
• The Kafka writer can track schema evolution through the Confluent schema registry

Monitoring and recovery

• A monitor page graphs events processed, application and server CPU, and memory across the cluster
• End-to-end lag is sampled every five seconds, retained 24 hours and charted in milliseconds
• Smart alerts to email, Slack, Teams or the web console, the last being the default
• JMX monitoring, server log access and file lineage are Striim Platform only, not Cloud
• Automatic failover and load rebalancing are unavailable on Cloud Enterprise, which is single-server

Where it runs and what it costs

• Striim Cloud is fully managed on AWS, Azure or Google Cloud, in Enterprise and Mission Critical tiers
• Private connectivity through Private Service Connect, Azure Private Link, VPN or a forwarding agent
• Striim Platform is self-managed on your own hardware or a cloud virtual machine
• A free Developer edition runs serverless with up to 25 million events a month
• No Kubernetes operator is published, and no region list appears on its pages

Pricing

Usage-basedFree Developer tier; Cloud at $0.60 a vCPU hour

The website is quote-only but the AWS Marketplace listing carries a real rate card for Striim Cloud Enterprise: $0.60 per vCPU hour, $0.10 a gigabyte in and out, and event credits at 80 per million events on standard adapters. Annual contracts run from $45,000 for 45,000 credits up to $243,000 for 243,000, with unit rates falling sharply at volume and unused credits rolling over. Overage is $1.00 a unit. The free Developer edition covers 25 million events a month. Striim Platform and Mission Critical are genuinely quote-only.

Vendor pricing page →

Demos and videos

• Striim overview, real-time integration to the cloud Product demo · YouTube
• SQL Server to Cloud SQL product demo Tutorial · YouTube

About Striim

Striim was founded by the core team from GoldenGate Software, including Ali Kutay, its former chairman and chief executive, alongside Steve Wilkes and Alok Pareek; no founding year is published on its own pages. It moved to downtown Palo Alto in February 2025 and keeps an office in Chennai. It is private and venture-backed, with a Series B expanded to $30m and a $50m Series C announced in March 2021 led by Goldman Sachs Growth Equity, with Summit Partners, Atlantic Bridge, Dell Ventures and Bosch Ventures; no total is published. A newsroom check found no funding, acquisition or ownership change in 2025 or 2026.

Palo Alto, California · striim.com