面向政府大规模协同任务的时空数据管理和分析研究
Published in NDBC, 2023
摘要: 政务大规模协同任务是数据规模大，需要多部门协同完成的复杂任务。现有流程主要通过线下方式执行，难以应对复杂的政务协同场景，且执行结果反馈缓慢。针对任务执行需求，以及对跨时间、空间的任务执行情况进行实时聚合分析面临的挑战，提出了时空任务树的概念和基于日志的树形结构统计值增量更新策略。时空任务树的设计满足政府大规模协同任务跨部门、跨层级执行场景，并支持灵活变更业务类型，解决了传统业务系统无法实现跨层级协同的问题。基于日志的树形结构统计值增量更新策略能够实现多维数据统计的秒级更新，大大减少了计算开销，提高了查询效率，解决了实时计算响应慢和离线统计资源耗费多，时效性差的问题。实验结果表明，时空任务树和统计策略具有可行性和高可用性。
Abstract: In recent years, digital government construction has become an important task for governments at all levels, and the security of collaborative government data has become a highly concerned issue. Due to the high mobility of government personnel, frequent addition and removal of permissions for personnel is required, and failure to update in a timely manner can lead to the risk of data leakage. The diversity of government affairs sce-narios requires high manpower costs to maintain consistency in permissions. Traditional access control models are difficult to adapt to the complex permission control needs in collaborative government office scenarios. In order to solve the problem of data permission control in collaborative government office scenarios, a government organiza-tion-based access control model (OBAC-G) is proposed. Firstly, the government organization structure is intro-duced into the access control model, the government organization tree is constructed,. Secondly, the government organization tree is further abstracted and mapped into an abstract hierarchical tree. Then, based on the vertical management administrative system of the government, a label system is constructed. Finally, three permission con-trol strategies are designed, and an administrator identity is introduced to achieve a full-process closed loop of organization batch authorization and semi-automatic authorization, improving the efficiency and security perfor-mance of permission configuration. Comparative analysis with other methods shows that this model is significantly better than other methods in government scenarios.
王璐璐、何华均、潘哲逸、鲍捷、郑宇. 面向政府大规模协同任务的时空数据管理和分析研究.[NDBC 2023].