Abstract:
In Big Data environments, reliability of data plays an important role to determine trustworthiness of the outcomes of an analysis. Big data provenance ensures the reliability of data by providing details about the origin and historical paths of data. In recent years, the preponderance of big data and its applications are increasingly using Apache Cassandra due to its high availability and linear scalability. In this paper, we present a data provenance framework for Key-Value Pair Databases using the concept of Zero-Information Loss Database (ZILD). A large volume of real-time social media data is fetched from the Twitter’s network through live streaming with the help of Twitter Streaming APIs, and then modelled in Apache Cassandra based on a Query-Driven approach. This framework provides efficient provenance capturing support for select, aggregate, update, and historical queries. We evaluate the performance of proposed framework in terms of provenance capturing and querying capabilities using appropriate query sets.