A System for Big Data Analytics over Diverse Data Processing Platforms

Abstract

Data analytics is at the core of any organization that wants to obtain measurable value from its growing data assets. Data analytic tasks may range from simple to extremely complex pipelines, such as data extraction, transformation and loading, online analytical processing, graph processing, and machine learning (ML). Following the dictum “one size does not fit all”, academia and industry have embarked on a race of developing data processing platforms for supporting all of these different tasks, e.g., DBMSs and MapReduce-like systems. Semantic completeness, high performance and scalability are key objectives of such platforms. While there have been major achievements in these objectives, users are still faced with many road-blocks. MOTIVATING EXAMPLE The first roadblock is that applications are tied to a single processing platform, making the migration of an application to new and more efficient platforms a difficult and costly task. As a result, the common practice is to re-implement an application on to...