计算机技术应用研究室知识库

The data volume in astronomical digital sky surveys has reached hundreds of PB per year. How to efficiently manage the massive astronomical data in explosive growth, how to achieve the integration on multi-wavelength observation data, and how to rapidly release the huge amounts of data generated by the larger telescopes are urgent problems needing to be addressed. The main purpose of this paper is to put forward an astronomical data warehouse architecture design to manage and publish massive astronomical data. In the proposed architecture data servers deal with the raw data obtained from telescopes based on a predefined integration strategy, and save the standardized metadata structure and information. We build a multi-level data storage model, to extract the different levels of data sets from the raw data. This multi-level data asset is formed by constructing an object-based database and data warehouse management system, where all data are centralized in a reliable storage. Catalogues with huge number of objects have not yet had an efficient cross matching algorithm to achieve data integration from different band observations. With the implementation of this data warehouse architecture, we were able to adopt a parallel computing model to evaluate the matching method of massive astronomical catalogue data. By improving the speed of matching, this method could potentially pave a way for the development of large-scale astronomical data of future larger telescopes and gain experience. 2014 ICIC International.