为提升主流关系型数据库在HTAP场景下的性能,本研究以PolarDB for PostgreSQL为对象,针对TPC-H基准测试进行了系统性的性能优化。通过深入分析数据导入与复杂查询过程中的性能瓶颈,本文构建并验证了一套多层次的协同优化策略。该策略涵...为提升主流关系型数据库在HTAP场景下的性能,本研究以PolarDB for PostgreSQL为对象,针对TPC-H基准测试进行了系统性的性能优化。通过深入分析数据导入与复杂查询过程中的性能瓶颈,本文构建并验证了一套多层次的协同优化策略。该策略涵盖逻辑层(SQL重构、约束延迟)、物理层(分区、聚簇、覆盖索引)以及系统与硬件层(内核参数调优、SIMD指令集加速)。实验结果表明,该系列优化措施效果显著:数据导入时间缩短81.14%,22条TPC-H查询的平均响应时间从超过10000 s降至369 s,底层数据校验性能提升86.90%。本研究证实,通过软硬件协同优化,传统关系型数据库在OLAP负载下的性能可获得显著的提升,为构建经济高效的HTAP解决方案提供了实证依据。展开更多
Population growth and rapid urbanization have caused an exponential increase in urban land demand in Ethiopia.The information required for land administration and management has been tremendously increasing as well.Th...Population growth and rapid urbanization have caused an exponential increase in urban land demand in Ethiopia.The information required for land administration and management has been tremendously increasing as well.Therefore,this paper describes the steps for a building database prototype for urban land information management system for the case of Ethiopia which results to develop an efficient and effective way of spatial and non-spatial urban land data management system.This research project will develop a central database system using a PostgreSQL software to retrieve,update and store the required data using a user-friendly graphical interface.System requirement and specification is devoted to a system design which incorporates the intended system behaviour using a use case diagram,logical design of the system using a class diagram.The design system is validated and tested via sample data which can store,retrieve,and update the required data as per the user requirement.Developing this type of urban land information management system,therefore,facilitates and increases the efficiency of land registration and related activities at the municipality level.Thence,citizens would be satisfied by the service delivery of the municipality.The land and real estate ownerships registration will also become automated.Furthermore,any spatial and non-spatial data related with the system can be stored in the database which will be ready for more analysis as to be able to use by policymakers,urban planners for the purpose of sustainable development of urban area infrastructural developments and land use policy.The research will lay ground for another task to develop CityGML 3D modelling for selected cities thereby contributes to spatial data infrastructure development to Ethiopia.展开更多
针对闪存缓冲区置换算法的性能验证基本采用仿真模拟的现状,提出了一种基于PostgreSQL的有说服力的性能验证方法,重点讨论了在PostgreSQL上扩展闪存缓冲区置换算法的方法和实现技术,并以CFLRU(clean first least recently used)和CCFLRU...针对闪存缓冲区置换算法的性能验证基本采用仿真模拟的现状,提出了一种基于PostgreSQL的有说服力的性能验证方法,重点讨论了在PostgreSQL上扩展闪存缓冲区置换算法的方法和实现技术,并以CFLRU(clean first least recently used)和CCFLRU(cold clean first least recently used)算法为例,给出了具体的扩展过程。然后以一块固态硬盘作为数据存储设备进行性能测试,测试结果证明了基于PostgreSQL的扩展方法在缓冲区置换算法性能验证上的有效性。展开更多
摘要为提升主流关系型数据库在HTAP场景下的性能,本研究以PolarDB for PostgreSQL为对象,针对TPC-H基准测试进行了系统性的性能优化。通过深入分析数据导入与复杂查询过程中的性能瓶颈,本文构建并验证了一套多层次的协同优化策略。该策略涵盖逻辑层(SQL重构、约束延迟)、物理层(分区、聚簇、覆盖索引)以及系统与硬件层(内核参数调优、SIMD指令集加速)。实验结果表明,该系列优化措施效果显著:数据导入时间缩短81.14%,22条TPC-H查询的平均响应时间从超过10000 s降至369 s,底层数据校验性能提升86.90%。本研究证实,通过软硬件协同优化,传统关系型数据库在OLAP负载下的性能可获得显著的提升,为构建经济高效的HTAP解决方案提供了实证依据。
摘要Population growth and rapid urbanization have caused an exponential increase in urban land demand in Ethiopia.The information required for land administration and management has been tremendously increasing as well.Therefore,this paper describes the steps for a building database prototype for urban land information management system for the case of Ethiopia which results to develop an efficient and effective way of spatial and non-spatial urban land data management system.This research project will develop a central database system using a PostgreSQL software to retrieve,update and store the required data using a user-friendly graphical interface.System requirement and specification is devoted to a system design which incorporates the intended system behaviour using a use case diagram,logical design of the system using a class diagram.The design system is validated and tested via sample data which can store,retrieve,and update the required data as per the user requirement.Developing this type of urban land information management system,therefore,facilitates and increases the efficiency of land registration and related activities at the municipality level.Thence,citizens would be satisfied by the service delivery of the municipality.The land and real estate ownerships registration will also become automated.Furthermore,any spatial and non-spatial data related with the system can be stored in the database which will be ready for more analysis as to be able to use by policymakers,urban planners for the purpose of sustainable development of urban area infrastructural developments and land use policy.The research will lay ground for another task to develop CityGML 3D modelling for selected cities thereby contributes to spatial data infrastructure development to Ethiopia.
摘要针对闪存缓冲区置换算法的性能验证基本采用仿真模拟的现状,提出了一种基于PostgreSQL的有说服力的性能验证方法,重点讨论了在PostgreSQL上扩展闪存缓冲区置换算法的方法和实现技术,并以CFLRU(clean first least recently used)和CCFLRU(cold clean first least recently used)算法为例,给出了具体的扩展过程。然后以一块固态硬盘作为数据存储设备进行性能测试,测试结果证明了基于PostgreSQL的扩展方法在缓冲区置换算法性能验证上的有效性。