一种高效的移动对象连续多范围查询处理框架

针对连续多范围查询处理,结合多核多线程技术和大容量内存技术,通过将移动对象和查询放在内存中处理,提出了一种基于多线程的连续多范围查询处理框架.该框架基于多核处理器平台采用多线程技术周期性地处理查询和移动对象的更新,并周期性地计算多范围查询的结果.提出了基于移动对象数据均匀划分的多线程连续多范围查询处理算法,该算法以为查询建立的格网索引为基础.给出了该索引的构建思想和更新算法.考虑到基于内存的算法受Cache访问性能影响,提出了基于空间填充曲线的移动对象存储优化方法.实验证明,基于多核平台的多线程处理能够高效地处理连续多范围查询,同时通过移动对象存储优化能够提高算法运行中Cache访问命中率,进而提高算法性能.

An Efficient Framework for Processing Continuous Multiple Range Queries on Moving Objects

To evaluate continuous multiple range queries,considering the rapid development of multicore multi-threading technology and huge main memory,a framework for processing continuous multiple range queries based on multi-threading is proposed,which keeps both the moving objects and queries in main memory.The framework can be executed on multi-core platform by using multithreading technology to periodically update the moving objects and queries,meanwhile periodically evaluate the multiple range queries.By using the grid index built for multiple queries,a multithreading continuous multiple range query algorithm is proposed based on uniform partition of the moving objects in memory.The building and updating algorithms for the grid index are also discussed.Considering that main memory algorithms are greatly affected by the cache access performance,a method is proposed to optimize the storage of the moving objects by using the space filling curves.Experimental results show that multi-threading processing of the continuous multiple range queries on multi-core platform is highly efficient.Meanwhile,through the storage optimization of the moving objects,the cache access hit-rates can be improved greatly,thus the overall performance of the algorithm is improved.