本代码库用于介绍Creek使用案例,扩展Creek的自定义函数,经验技术交流等
Creek是一个轻量级的通用流式计算框架,作业内存消耗<10MB,运行时零依赖。Creek提供与Apache Flink完全兼容的SQL接口,采用申明式的作业定义,在云端静态编译,以可执行文件方式下发。
Creek在线作业编辑生成,请访问: http://creek.baidubce.com/
This Repo demos use case of Creek, as well as sharing UDF, and also serves as a place for communication.
Creek is a super lightweight general purpose stream processing framework, it's implemented in Go. It provides Apache Flink compatible SQL interface, user declares jobs in a JSON document, no coding is required.
Creek takes less than 10MB of memory, hence it can run on most of the resource constrained devices, e.g. edge devices.
Jobs are statically linked to executables, so it has no dependency during runtime, therefore deploying it is as simple as download and launch.
To build a Creek job, refer to http://creek.baidubce.com/