Twitter Heron: Towards Extensible Streaming Engines
Twitter’s data centers process billions of events per day the instant the data is generated. To achieve real-time performance, Twitter has developed Heron, a streaming engine that provides unparalleled performance at large scale. Heron has been recently open-sourced and thus is now accessible to various other organizations. In this paper, we discuss the challenges we faced when transforming Heron from a system tailored for Twitter’s
applications and software stack to a system that efficiently handles applications with diverse characteristics on top of various Big Data platforms. Overcoming these challenges required a careful design of the system using an extensible, modular architecture which provides flexibility to adapt to various environments and applications. Further, we describe the various optimizations that allow us to gain this flexibility without sacrificing performance. Finally, we experimentally show the benefits of Heron’s modular architecture.