摘要
随着车联网与大数据技术的发展,车辆管理服务平台架构由传统的分散式、本地化,走向集中式、扁平化,平台端资源消耗越来越大,如何实现海量车载终端数据的高并发实时采集是一个亟待解决的问题。针对该问题,提出基于Boost.Asio网络通信库的解决方案。该方案以前后端分离的方式,将数据采集与数据解析分开实现,利用Boost.Asio的前摄器模式实现高并发的数据采集,利用Kafka消息队列提高系统可扩展性,并结合线程池及智能指针技术,对传统数据采集系统中数据量大、种类多带来的资源占用高问题进行了改进。实验结果表明,该系统在有大量连接的情况下,内存占用较少,连接稳定,数据无丢失,保证了数据采集的质量和可靠性。
With the development of car networking and big data technologies,the platform of vehicle management service platform has been decentralized and localized from traditional to centralized and flat. Platform-side resource consumption is growing. How to realize the high concurrency real-time data acquisition of the huge car terminal data is an urgent problem to be solved. To solve this problem,a solution based on Boost. Asio network communication library was proposed. The scheme achieved data acquisition and data analysis separately by separating front-end and back-end. The Boost. Asio's forearm model was used to achieve high concurrent data acquisition. It improved system scalability with Kafka Message Queuing. Combined with the thread pool and smart pointer technology,the traditional data acquisition system had been improved in terms of large amount of data and high occupancy of resources. Experimental results showed that the system had less memory occupation,stable connection and no data loss when there were a large number of connections,which ensured the quality and reliability of data acquisition.
出处
《计算机应用与软件》
北大核心
2018年第2期248-255,共8页
Computer Applications and Software
基金
国家自然科学基金项目(61273323)
