Kafka のコストを 25% 以上削減 | Confluent コスト削減チャレンジに参加
With a rich ecosystem and support for multiple languages, it’s no surprise that Protobuf has emerged as a challenger to Avro’s crown as the de-facto serialization format for Kafka. Helped by first class support from Confluent, RedHat and others, Protobuf has finally arrived as viable choice for enterprise wide use cases.
During this talk we will tackle how we have used Protobuf successfully with Kafka: from clients to connectors; streams to schema registry; and gitops to governance. We will go over our learnings, including how we have improved the developer experience through the use of linting and early breaking change detection.
Expect to leave this talk knowing more about Protobuf and how it is supported across the Kafka ecosystem. We will cover thorny topics such as field presence, reusability, the value of a registry and when schema-less is the right answer. Finally, we will share the pitfalls and challenges, how we’ve made Protobuf work seamlessly for us at scale.