Lorenzo Alberton has a new post with an update about a library he's been working on to interface with Apache's Kafka system, "a persistent, distributed, high-throughput publish-subscribe messaging system".
Over a year ago I sort of reverse-engineered the protocol, and published a first rudimental library to produce and consume messages with version 0.05. Since then, the Kafka project has evolved a lot, moving from the LinkedIN repositories to the Apache Incubator, gaining traction, committers, features. The protocol has changed slightly too, so the old library doesn't work anymore with the new version (0.6+).
The library has gotten lots of new features in this update including gzip compression support, custom exception handling and better connection handling. He includes some sample code in the post showing how to create both the Producer and Consumer for the messages (and one for working with Zookeeper).