the current #Kafka protocol #pubsub #protocols
on 02021-01-13the #Kafka protocol, long ago. #pubsub #protocols
on 02021-01-13detailed #performance comparison of #Kafka and #RabbitMQ (and #Redis)
on 02017-06-18another comparison of #0MQ vs. #nanomsg and other things including #Kafka and #RabbitMQ and even #Redis
on 02017-06-18a #0MQ vs. #Kafka comparison, but this one mostly focused on their “Quasar” actors programming framework
on 02017-06-18#Kafka is “less like a queue system and more like a circular buffer that scales as much as your disk on your cluster” and “doesn’t have message acknowledgements, it assumes the consumer tracks of what’s been consumed so far”, while #RabbitMQ “is broker-centric, focused around delivery guarantees between producers and consumers, with transient preferred over durable messages;” while Kafka handles 10+ million messages per second at LinkedIn, RabbitMQ struggles to exceed 1 million.
on 02017-06-18The rationale for #Kafka’s design
on 02017-06-18Auth0 Webtasks (?) switched their log aggregation from #Kafka to #0MQ; apparently Kafka is much faster than #RabbitMQ
on 02017-06-18Hortonworks’ #Kafka #introduction: “a fast, scalable, durable, and fault-tolerant publish-subscribe messaging system”
on 02017-06-18Cloudera’s #Kafka #introduction: “a distributed publish-subscribe messaging system that is designed to be fast, scalable, and durable.…Messages are simply byte arrays…It is possible to attach a key to each message”
on 02017-06-18Jay Kreps describes how he invented #Kafka and how log-centric computing could or should work.
on 02017-06-14