7 Reasons We Chose Apache Pulsar over Apache Kafka

Kesque, an AI company, has chosen Apache Pulsar over Apache Kafka for building its messaging infrastructure service due to several reasons. Firstly, Apache Pulsar is a two-in-one solution that can handle high-rate real-time use cases like Kafka and also supports standard message queuing patterns. Secondly, unlike Kafka, Apache Pulsar doesn't require partitions for all topics, making it simpler to use in some scenarios. Thirdly, Apache Pulsar avoids the problem of copying large logs by breaking them into segments, which are distributed across multiple servers using Apache BookKeeper as its storage layer. This makes failure scenarios easier to handle and scaling out a breeze. Fourthly, Apache Pulsar brokers are stateless, allowing for quick start-up times and easy scalability. Lastly, geo-replication is a first-class feature in Pulsar, making it suitable for globally distributed applications or disaster recovery scenarios. Additionally, benchmark tests have shown that Pulsar delivers higher throughput along with lower and more consistent latency compared to Kafka.