Holik Martin, Horvath Tomas, Oujezsky Vaclav, Munster Petr, Tomasov Adrian, Valach Sobeslav
Department of Telecommunication, Brno University of Technology, Technicka 12, 616 00 Brno, Czech Republic.
Sensors (Basel). 2020 Oct 30;20(21):6208. doi: 10.3390/s20216208.
This work is focused on creating an open-source software-based solution for monitoring traffic transmitted through gigabit passive optical network. In this case, the data are captured by the field-programmable gate array (FPGA) card and reassembled using parsing software from a passive optical network built on the International Telecommunication Unit telecommunication section (ITU-T) G.984 gigabit-capable passive optical network GPON recommendation. Then, the captured frames are converted by suitable software into GPON frames, which will be further processed for analysis. Due to the high transfer rate of GPON recommendations, the work describes the issue of writing to the Mongo database system. In order to achieve the best possible results and minimal loss of transmitted frames, a series of tests were performed. The proposed test scenarios are based on different database writing approaches and are implemented in the Python and C# programming languages. Based on our results, it has been shown that the high processing speed is too high for Python processing. Critical operations must be implemented in the C# programming language. Due to rapid application development, Python can only be used for noncritical time-consuming data processing operations.
这项工作专注于创建一个基于开源软件的解决方案，用于监控通过千兆无源光网络传输的流量。在这种情况下，数据由现场可编程门阵列（FPGA）卡捕获，并使用解析软件从基于国际电信联盟电信部门（ITU-T）G.984千兆无源光网络GPON建议构建的无源光网络中重新组装。然后，捕获的帧由合适的软件转换为GPON帧，这些帧将进一步处理以进行分析。由于GPON建议的高传输速率，该工作描述了写入Mongo数据库系统的问题。为了获得尽可能好的结果并使传输帧的损失最小化，进行了一系列测试。所提出的测试场景基于不同的数据库写入方法，并使用Python和C#编程语言实现。根据我们的结果，已经表明对于Python处理来说，高处理速度过高。关键操作必须用C#编程语言实现。由于应用程序开发迅速，Python只能用于非关键的耗时数据处理操作。
