近年来面向存算分离场景的大数据应用获得极大的关注，Alluxio作为首个数据编排系统，解决了数据驱动应用与存储系统之间的数据访问困难的问题，并在提升系统性能、降低运营成本方面有显著效果，但目前缺少一套性能测试工具，可对数据编排软件各方面性能进行定量化描述。文章针对数据编排系统Alluxio设计并实现了一套性能测试工具，集成了主流大数据软件、基准测试以及监控系统，对Alluxio的性能进行全方面的测试，即能比较不同版本、不同运行模式下的Alluxio以及其配套软件的性能差别，同时为宣传产品竞争力提供数据支撑。测试工具及相关文档已发布至Alluxio开源社区，为研发人员提供灵活的测试方法和监控方案。

In recent years, big data applications for storage-computing separation scenarios have gained significant traction. As the first data orchestration system, Alluxio solves the problem of difficult data access between data-driven applications and storage systems, and improves system performance while reducing operating costs. There is a significant effect in terms of cost, but currently there is a lack of a suite of performance testing tools that can quantitatively describe the performance of all aspects of data orchestration software. This article designs and implements a suite of performance testing tools for the data orchestration system Alluxio, which integrates mainstream big data frameworks, benchmark tests, and monitoring systems to test the performance of Alluxio in all aspects. That is, it can compare the performance differences of Alluxio and its supporting software in different versions and different operating modes, and provide data support for promoting product competitiveness. The test tools and related documents have been released to the Alluxio open-source community, providing developers with flexible testing methods and monitoring solutions.