This paper presents a novel transversal, agnostic-infrastructure, and generic processing model to build environmental big data services in the cloud. Transversality is used for building processing structures (PS) by reusing/coupling multiple existent software for processing environmental monitoring, climate, and earth observation data, even in execution time, with datasets available in cloud-based repositories. Infrastructure-agnosticism is used for deploying/executing PSs on/in edge, fog, and/or cloud. Genericity is used to embed analytic, merging information, machine learning, and statistic micro-services into PSs for automatically and transparently converting PSs into big data services to support decision-making procedures. A prototype was developed for conducting case studies based on the data climate classification, earth observation products, and making predictions of air data pollution by merging different monitoring climate data sources. The experimental evaluation revealed the efficacy and flexibility of this model to create complex environmental big data services.

本文提出了一种新颖的横向、不可知基础设施和通用处理模型，用于在云中构建环境大数据服务。横向用于构建处理结构 (PS)，通过重用/耦合多个现有软件来处理环境监测、气候和地球观测数据，甚至在执行时间，数据集在基于云的存储库中可用。基础设施不可知论用于在边缘、雾和/或云上/中部署/执行 PS。通用性用于将分析、合并信息、机器学习和统计微服务嵌入到 PS 中，以自动透明地将 PS 转换为大数据服务，以支持决策程序。开发了一个原型，用于基于数据气候分类、地球观测产品、并通过合并不同的监测气候数据源来预测空气数据污染。实验评估揭示了该模型在创建复杂环境大数据服务方面的有效性和灵活性。