CAREER: Tracing the Provenance and Flow of Data

该研究项目的目标是研究技术和开发有效的工具，用于追踪相互关联的数据库网络中的数据来源和流动，其中每个数据库都可能不断发展。本文主要研究了以下三个方面的问题：（1）基于约束的关系/XML数据的出处推理。数据库或数据库版本之间的关系通常可以通过高级抽象（如约束）来描述。因此，跟踪一段数据的来源或流涉及通过约束推理一段数据的移动。(2)基于出处的注释管理系统。该系统提供了一种计算出处的替代方法。它通过在数据转换时将出处沿着作为注释来“急切地”这样做。(3)基于关键限制的归档工具包。提供了一组实用程序，用于以语义上有意义的方式有效地归档和管理不断变化的数据库。档案是这样的，一段数据的出处可以很容易地跟踪。该项目的成果在许多领域，特别是在生物信息学等科学数据库中具有潜在的应用价值。在适用的情况下，将对该项目的结果进行原型制作，并对合成和真实的数据集进行实验评价，以评估结果的有效性。结果也将被集成到一个先进的数据库主题课程的数据出处，将开发和教授在加州大学圣克鲁斯。项目网站http://www.cs.ucsc.edu/~wctan/provenance/将用于广泛传播所有课程材料、出版物和项目产生的软件。

The goal of this research project is to investigate techniques and develop effective tools for tracing the provenance and flow of data in a network of inter-related databases where each database may be constantly evolving. The following three key areas are investigated: (1) Reasoning about provenance for relational/XML data through constraints. The relationships between databases or versions of databases can often be described by high-level abstractions such as constraints. Hence, tracing the provenance or flow of a piece of data involves reasoning about the movement of a piece of data through constraints. (2) An annotation management system based on provenance. This system provides an alternative method for computing provenance. It does so "eagerly" by carrying provenance along as annotations when data is transformed. (3) An archiving toolkit based on key constraints. A set of utilities for efficiently archiving and managing evolving databases in a semantically meaningful way is provided. The archive is such that the provenance of a piece of data can be easily traced. The results of this project have potential applications in many domains, in particular in scientific databases such as bioinformatics. Whenever applicable, results of this project will be prototyped and experimental evaluations will be conducted on both synthetic and real datasets in order to evaluate the effectiveness of the results. The results will also be integrated into an advanced database topics course on data provenance that will be developed and taught at UC Santa Cruz. Project website http://www.cs.ucsc.edu/~wctan/provenance/ will be used for wide dissemination of all course materials, publications, and software that result from this project.