Home

Introduction

Welcome to OSPBench (Open Stream Processing Benchmark) - a benchmarking suite for distributed stream processing frameworks.

This wiki contains documentation for developers and researchers working on benchmarking and testing distributed stream processing frameworks.

Currently, this project includes four frameworks: Apache Flink, Apache Kafka Streams, Apache Spark: Spark Streaming and Structured Streaming.

This suite can be used for benchmarking latency, peak sustainable throughput, peak throughput, scalability and fault tolerance.

Behavior can be analyzed for different pipeline complexities, data characteristics and configuration settings.

Some uses of this code base can be for benchmarking:

• common stream operations between frameworks and within frameworks
• differences in bottlenecks between frameworks as pipelines become more complex
• latency of different pipelines and implementations
• throughput
• resilience against data bursts
• scalability and resource efficiency
• fault tolerance under different settings and configurations
• speed of recovery
• aspects of state management: influence on recovery, influence on performance,...
• ...

Table of contents

Quick navigation:

• Benchmark
  • Data
  • Processing Pipelines
  • Workloads
• Architecture
• Local deployment
• Deployment on AWS with DC/OS
• Metrics
• Frameworks
• Results
• Data Stream Generator
• Metrics Exporter
• Output Consumer
• Evaluation Suite
• References, Publications and Talks

Questions or need help?

Do you wish to use this project or extend it? The fastest way to contact me is through:

• LinkedIn: giselle-van-dongen
• Email: [email protected]