适用于广域测量系统的实时数据并发访问同步算法

发布时间：2018-11-23 11:01

【摘要】：由于广域测量系统(wide area measurement system,WAMS)海量实时数据的大规模、高负荷并发访问,其采用的同步机制在很大程度上约束了系统的效率。针对于目前普遍使用的整体锁定机制带来的由于访问串行化导致的效率低下问题,提出了粒度可控并发访问同步算法(controllablegranularity concurrency synchronization,CGCS),该方法使用控制标志位建立基于子集超集依赖的锁定条件和先进先出等待队列,并可控锁定级别,实现对实时数据访问的互斥粒度粗细的任意控制,同时作用于整体文件、表、元组,进而任务可以最大化地并发执行。通过实验,给出了系统的并发处理能力和IO响应能力的测试过程和结果,证明了CGCS算法在并发高、访问散的情况下能充分发挥CPU的并行处理能力,使WAMS系统的效率得到大幅提升。
[Abstract]:Due to the large scale and high load concurrent access of the massive real-time data of the wide-area measurement system (wide area measurement system,WAMS), the synchronization mechanism used in the system to a large extent restricts the efficiency of the system. Aiming at the inefficient problem caused by the serialization of access caused by the universal locking mechanism, a granularity controllable concurrent access synchronization algorithm (controllablegranularity concurrency synchronization,CGCS) is proposed. The method uses the control flag bit to establish the locking condition based on the subset superset dependency and the first-in-first-out wait queue, and controls the locking level, realizes arbitrary control of the mutex granularity of real-time data access, and acts on the whole file and table. Tuples, which in turn maximize concurrent execution of tasks. Through experiments, the test process and results of concurrent processing ability and IO response ability of the system are given. It is proved that the CGCS algorithm can give full play to the parallel processing ability of CPU in the case of high concurrency and scattered access. The efficiency of WAMS system is greatly improved.
【作者单位】：华北电力大学控制与计算机工程学院;
【基金】：高等学校博士学科点专项科研基金课题(20120036120003) 中央高校基本科研业务费专项资金项目(12ZP09)~~
【分类号】：TM76

【参考文献】