1. 农业大数据基础设施开发的参考模型方法
- Author
-
Zhao, Z., Liao, X., Wang, X., Ruan, C., Zhu, Y., Feng, D., and System and Network Engineering (IVI, FNWI)
- Abstract
Big data infrastructures provide services for the management of data over the course of their lifecycle, and offer users the ability to effectively discover and access data for different application purposes. These emerging infrastructures essentially enable system-level data-centric research; third-party innovation, however, often requires data from different sources. The construction of big data infrastructures faces important interoperability challenges arising from the diverse nature of data acquisition, annotation, and identification performed in different research domains. Moreover, the evolution of different infrastructures is often driven by the specific interests of researchers, in their respective domains, and the constraints of legacy technology. The ENVRI Reference Model (ENVRI RM) is an output of the EU H2020 ENVRI and ENVRI PLUS project, targeting the aforementioned challenges in the context of environmental sciences by modeling environmental research infrastructures with a multi-viewpoints framework; these viewpoints include science, information, computation, engineering, and technology. Each viewpoint describes concrete aspects of a system definition and forms a mechanism to improve the interoperability across the whole system as well as alignment with existing legacy systems. The challenges encountered in the Shanghai Agricultural Big Data Infrastructures construction work are similar to those detected in the ENVRI RM, which provides an ideal place to test the generalizability of the ENVRI RM to other domains. Using the ENVRI RM as a reference, this paper presents an Agricultural Reference Model, which includes the five aforementioned viewpoints, but with consideration of the specifics of the agricultural domain, to address the problems encountered in revising and upgrading the Shanghai Agricultural Big Data Infrastructures. Two use cases are introduced to demonstrate its effectiveness. One is to improve the requirement engineering procedure with the community and role context captured using the Agricultural Reference Model. The other is to upgrade the large volume of existing systems to increase interconnections via the interoperability mechanisms provided by the Agricultural Reference Model.
- Published
- 2019