Flink and Kafka are widely used in data channel scenarios inside Didi, including data warehouse, log diagnosis, data analysis and other scenarios. In the process of the system evolvement, we have accumulated a set of mature log collection, delivery, integrity judgment mechanism, as well as experiences in performance and cost optimization. In this talk, we will introduce our optimization and practical experience in such scenarios.

Speakers:

Xue Kang: Head of real-time computing and data channel at Drip, dedicated to building a powerful and easy-to-use real-time data analysis system and a stable and low-cost data transmission service.