Short Intro to Flume
The Flume ☞ user guide is 70 screen long, so use these slides and the Flume and Scribe comparison as short intros:
Notes:
- What problems are solved by Flume?
- data collection in all formats
- flexible reliability guarantees allowing careful performance tunning
- quick iteration on new collection strategies
- Flume is build around teh concept of flows corresponding to a type of data source and composed from chained nodes
- A Flume node receives data from a source, optionally processes it using 1 or more decorators and outputs it via a sink
- Nodes receiving data are called agents, nodes writing data are called collectors
Original title and link for this post: Short Intro to Flume (published on the NoSQL blog: myNoSQL)
