Cloud Stream Service

Cloud Stream Service (CS) provides full-stack capabilities for processing streaming data in real time. Compatible with Apache Flink and Spark APIs, CS fully hosts computing clusters, so you can run StreamSQL or user-defined jobs without learning any programming skills.

Pay per use, SPU unit price $0.0679 USD/hour

Learn more
Product Advantages
  • Easy to Use

    CS allows you to implement business logic by using StreamSQL statements. You only need to perform streaming data analysis without the need to manage clusters and learn programming skills.

  • Support of Exclusive Clusters

    You can run your jobs in a shared cluster or exclusive cluster. Exclusive clusters are physically isolated from shared clusters and other tenants' clusters. You can also manage the quota of exclusive clusters.

  • Pay per Use

    The service is priced based on the used SPU resources and the service duration (by second). An SPU contains one core and 4 GB memory.

  • High Throughput, Low Latency

    CS uses the Dataflow model of Apache Flink, a real-time computing framework. High-performance computing resources are used to consume data from your created Kafka, DMS Kafka, and MRS Kafka clusters. A single SPU processes 1,000 to 50,000 messages per second.

Application Scenarios
  • Real-time Stream Analysis

  • IoT

Real-time Stream Analysis

Real-time Stream Analysis

Real-time stream analysis features ease of use, low latency, and high throughput. It can be achieved based on StreamSQL and user-defined jobs.


Easy to Use

Supports online StreamSQL statement editing and provides abundant SQL functions to meet complex service requirements.

Fully Managed

Computing clusters are fully managed by CS, enabling you to focus on stream analysis.

Pay per Use

SPU resources specified during job creation are charged by duration (unit: second).



IoT or edge devices upload data to DIS. CS reads data from DIS, analyzes data (including fault detection and counter warning), and makes the analysis result persistent or reports alarms in real time.


Abundant IoT SQL Functions

Provides common IoT functions such as area, yaw, and distance detection functions.

High Throughput, Low Latency

Leverages Apache Flink to achieve a complete real-time computing framework.

Secure Isolation

Isolates tenants from each other to ensure data security.

Basic Functions

  • StreamSQL Functions

    SQL is used to express the business logic. Aggregate functions, such as WINDOW and JOIN are supported.

  • High Throughput, Low Latency

    CS supports the backpressure mechanism, high throughput, and millisecond-level latency.

StreamSQL Functions

High Throughput, Low Latency

  • Rock-solid Security

    Security protection mechanisms for tenants ensure operation security.

  • Distributed Real-time Computing

    Large-scale cluster computing and cluster auto scaling significantly reduce the cost.

Rock-solid Security

Distributed Real-time Computing

Usage Guidelines

Provides Scalable, On-demand Computing Resources

Learn More