Storm [1] and Spark [2] are two other distributed stream processing libraries in this family. Lots of good work being done in this area; batch-processing a la Hadoop is simply unnecessary for a lot of incredibly useful algorithms.
As is Streams [1], which I work on. There are many research papers on the project which use the research name System S to differentiate active research from the product.
5 comments
[ 0.24 ms ] story [ 10.9 ms ] threadA benchmark of how it performs against the other solutions would be great!
[1] https://github.com/nathanmarz/storm [2] http://www.spark-project.org/
[1] http://www-01.ibm.com/software/data/infosphere/streams/