Abstract
Parallel processing is one of the most significant trends in data processing. Several parallelization frameworks have been created to provide more advanced programming environment than the low-level services provided by the operating system.
Nevertheless, developing parallel programs requires significant knowledge about synchronization, cache hierarchy and a lot of other technical details. Bobox is a parallel processing framework designed specifically for processing of large amounts of data.
It provides easier to use interface and possibly better performance for a certain class of problems - data-intensive computations based on a non-linear pipeline. Using the framework makes development of parallel data processing tools much simpler without sacrificing efficiency of the computation.