Towards an Optimized Big Data Processing System

Provided by: Delft University of Technology
Topic: Big Data
Format: PDF
To perform fast and inexpensive big data analytics, researchers use a processing system represented by a stack of frameworks for data storage, data processing, and data manipulation deployed over a large distributed system. In the context of the data explosion phenomenon, existing performance models for MapReduce are applicable for specific production workloads, but are yet to reveal the real capabilities of the processing system under heavy workloads that process tens of terabytes of data. Therefore, the research is to optimize the MapReduce processing system for processing terabytes of data.

Find By Topic