Hello cafe, I am happy to announce task-distribution, a framework written in order to distribute and execute tasks as fast as possible on a large data set. Its accentuation is that it supports distributing new logic to other nodes from within the program itself, thus avoiding both restarting slave node processes and a separated plugin-management infrastructure. An example use case is running continuously changing analytics logic on a given data set stored in HDFS. Avalailable on hackage and github: https://github.com/michaxm/task-distribution Cheers, Axel