Model transformations of MapReduce Design Patterns for automatic development and verification

摘要

Mapping MapReduce frameworks to Cloud Architecture became a must in last years because of the need of managing large data sets and Big Data in fast, reliable (and as cheap as possible) way. Scientific Literature proposes many works about new architectures, frameworks and algorithms improving and optimizing performances while performing MapReduce operations. Anyway, MapReduce framework is only the starting point for building a plethora of services based on different analyses. This is the reason for recent application of Design Patterns methodologies to develop MapReduce applications. Here we propose a Model Driven Engineering methodology to design, verify and develop MapReduce applications on Cloud Systems. The methodology is driven by MapReduce Design Patterns and is used to analyse soundness and reliability of services based on MapReduce from early design stage to runtime.