It could be used for ingesting data that’s destined for a Hadoop install, for example, or for feeding stream processing services based on Apache Samza (which also came out of LinkedIn).