向剑平, 乔少杰, 胡剑. WMB*:一种提高大数据上软件执行效率改进算法[J]. 内江师范学院学报, 2012, (12): 24-28.
    引用本文: 向剑平, 乔少杰, 胡剑. WMB*:一种提高大数据上软件执行效率改进算法[J]. 内江师范学院学报, 2012, (12): 24-28.
    XIANG Jian ping, QIAO Shao jie, HU Jian. WMB*:An Improved Algorithm to Enhance the Software Enforcement Efficiency in Handling Large Data[J]. Journal of Neijiang Normal University, 2012, (12): 24-28.
    Citation: XIANG Jian ping, QIAO Shao jie, HU Jian. WMB*:An Improved Algorithm to Enhance the Software Enforcement Efficiency in Handling Large Data[J]. Journal of Neijiang Normal University, 2012, (12): 24-28.

    WMB*:一种提高大数据上软件执行效率改进算法

    WMB*:An Improved Algorithm to Enhance the Software Enforcement Efficiency in Handling Large Data

    • 摘要: 为了提高软件的执行效能及提高WMB(WebSphere Message Broker)上大数据消息数据处理的速度,首先定义了单笔柜面交易、硬件标尺、程序执行效率等概念.建立了企业服务总线上的软件执行高效算法(称为WMB*),算法主要解决了在WMB上提高各应用级系统对大数据消息数据及通讯协议的处理和转换速度,以达到提高软件的执行效率.在银行数据集上对不改变大数据消息数据结构和改变大数据消息数据结构的情况做了大量的实验,结果表明,在WMB上,用ESQL语言比JAVA语言在处理消息数据的速度快0.1个数量级,而在CPU占用率(ms/msg)上,用ESQL语言比JAVA语言低得多.WMB*算法适合在大数据上对银行交易数据进行挖掘.

       

      Abstract: In this study, an algorithm on ESB (Enterprise Service Bus) called WMB* is proposed to improve the execution efficiency of software and the speed of message processing in large data on WMB (Websphere Message Broker). The new concepts including single counter transaction, hardware scale and efficiency of program execution are defined. And the new algorithm successfully enhances the processing capacity and the transfer speed of large data on WMB so as to improve the execution efficiency of software. Extensive experiments are conducted in the two cases of changing or not changing the data structure of messages in large data on bank data sets. The experimental results show that the speed of message processing using ESQL is faster than that using JAVA with an improvement of 0.1 order of magnitude. The algorithm using ESQL can achieve lower usage of CPU than that using JAVA. The WMB* algorithm is verified to be suitable for mining massive data in bank transaction.

       

    /

    返回文章
    返回