Time and Location: 5/8/2019, Wed., 10:30 a.m., CS Conference Room 206

 

Speaker: Mr. Wei Zhang

 

Title: Exploring Metadata Search Primitives for Scientific Data Management

 

Abstract: Scientific experiments and observations tend to store large amount of datasets in scientific data formats. To rapidly sift through the datasets and to locate the target dataset, metadata is commonly used as a clue to achieve efficient metadata search. While building in-memory metadata index is a beneficial practice for such purpose, the choice of in-memory data structures remains nebulous in the context of scientific data management, due to the lack of investigation on metadata, metadata queries and corresponding data structures. In this study, we explore the metadata search primitives in the context of scientific data management. We conduct our investigation on real-world scientific dataset and we explore the characteristics of the metadata in the dataset. We infer possible metadata queries based on our discovery of the metadata characteristics and we systematically evaluate different data structures for different data types of metadata attributes. We conclude our discovery analytically and quantitively at the end of this talk.