A Hierarchical Framework for Provenance Based on Fragmentation and Uncertainty

In the recent past, provenance – a research field that can be used to determine the origin and derivation history of data – has gained much attention [12]. Additionally, provenance is an important field for validating computation results. It covers coarse grained data as well as very fine grained information showing details of the implementation. Moreover, it is highly related to different topics such as causality [25]. Unfortunately, there is currently no global framework covering existing approaches and addressing the versatile characteristics of provenance. In this paper, we suggest a hierarchical framework for provenance based on its most general and common characteristics w.r.t. to the current state of the art. In fact, the framework contains different layers of abstraction. We discuss the use and limitations of all layers by means of existing models and formalisms that lay the foundation for each layer. Additionally, we explain the relationship between these layers and analyze how existing approaches interact with our framework.