Design and Implementation of Metadata System in PetaShare

摘要

As the size of scientific and commercial datasets grows, it becomes imperative that an expressive metadata framework to be developed to facilitate access to the semantics of the data. However, the drive to enhance expressiveness and maximize performance often turns out to be contradictory. Hence, the design of metadata framework needs to take into consideration many factors in order to achieve balance between expressive power and performance. In this paper, we discuss the metadata framework developed for PetaShare and explain the design and implementation decisions we made in the process.