英文:
KafkaStream Vs Flink
问题
我已经使用Flink将数据从源发送到目标。
我的Flink应用程序从Kafka消费数据并发送到目标。
目标也是Kafka主题，具有不同的主题名称。
Flink仅用于传递数据，没有任何业务逻辑。
在这种情况下，我认为将Flink更改为Kafka Stream将增加吞吐量。因为除了从源到汇的传递数据之外，Flink没有其他贡献。而且我的源和汇都使用Kafka，所以我认为在传递数据的情况下Kafka Streams会更快。
如果您对我的问题有任何意见，我将不胜感激。
谢谢。
英文:
I have used the Flink for sending data from source to sink.
My flink app consumes the data from Kafka and send to the destination.
The destination is also kafka topic which has a different topic name.
The Flink is only used for delivering purpose without having any business logic.
In this case, I think that changing the flink to Kafka Stream will increase the throughput. Because the flink has no contribution except for delivering data from source to sink. Also my source and sink uses the kafka so I think that kafka streams will be faster in case of the delivery data.
I would appreciate if you could give you any opinion for my question.
Thanks.
答案1
得分: 2
没有保证哪一个会比另一个更快。您仍然需要进行JVM和网络调优。
任何一个都可以工作，但Kafka Streams的限制是数据必须保留在同一个Kafka集群中。Flink没有这样的限制。
或者，您可以简单地使用MirrorMaker来在不同集群的Kafka主题之间传输数据。
英文:
There's no guarantee one will be faster than the other. You still need to do JVM and network tuning.
Either will work, but the limitation of Kafka Streams is that the data must remain in the same Kafka cluster. Flink has no such limitation.
Or you can simply use MirrorMaker for moving data between Kafka topics of different clusters.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论