基于LSTM-DT模型的低碳工业互联网网络安全态势平台  

Low-Carbon Industrial Internet Network Security Situation Platform Based on LSTM-DT Model

作　　者：李立伟 金鑫城 LI Liwei;JIN Xincheng(China Industrial Cyber Security and Development Research Center,Beijing 100043,China;State Grid Beijing Yizhuang Power Supply Company Daxing District,Beijing 100176,China)

机构地区：[1]国家工业信息安全发展研究中心,北京100043 [2]国网北京亦庄供电公司,北京100176

出　　处：《信息与电脑》2024年第6期71-73,共3页Information & Computer

摘　　要：低碳工业互联网网络安全态势平台的网络拓扑结构具有一定的限制性,导致节点之间的通信受到约束,使得平台吞吐量较小,无法满足相应的吞吐量需求。为此,提出基于长短时记忆-决策树(Long Short-Term Memory-DecisionTree,LSTM-DT)模型的低碳工业互联网网络安全态势平台设计研究。将GF-PR6K可编程控制面板作为硬件装置;利用长短期记忆网络,在记忆当前时刻低碳工业互联网网络状态信息的基础上,在下一时刻有选择地传递或遗忘低碳工业互联网网络的隐藏状态,最大限度降低低碳工业互联网网络安全态势分析阶段的冗余计算量,利用DT对低碳工业互联网网络状态序列信息进行综合。在测试结果中,设计平台吞吐量不仅呈现出了随着测试互联网网络构成及规模的增加逐渐增大的特点,且对应的涨幅明显,当节点之间通信距离由15 m增加至50 m时,对应的吞吐量增幅仅为1.00 kb。The network topology of the network security posture platform of low-carbon industrial Internet has certain limitations,resulting in restricted communication between nodes,resulting in small throughput of the plaform and unable to meet the corresponding throughput requirements.Therefore,the low-carbon industrial Internet network security situation platform design and research based on LSTM-DT model is proposed.The GF-PR6K programmable control panel is used as the hardware device of this paper.Using long and short-term memory network,the memory of the current moment low carbon industry on the basis of the Internet network information,in the next moment selectively transfer or forget the hidden state of low carbon industrial Internet network,minimize low carbon industry Internet network security situation analysis stage of redundant calculation,using DT of low carbon industry Internet network state sequence information.In the test results,the throughput of the design platform not only showed the characteristics of gradually increasing with the increase of the composition and scale of the test Internet network,but also the corresponding increase was obvious.When the communication distance between nodes increased from 15 m to 50 m,the corresponding throughput increase was only 1.00 kbit.

关 键 词：长短时记忆-决策树(LSTM-DT模型) 低碳工业互联网 网络安全态势 GF-PR6K可编程控制面板 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]