被動分簇下云服務器通信串口故障數據識別

摘" 要： 由于云服務器通信網中數據流量龐大且復雜，同時受到網絡結構和配置的多樣性以及動態變化的影響，傳統的主動探測或人工分析方法難以準確識別漏洞弧段，導致故障數據識別的準確性和效率受到限制。因此，研究一種基于被動分簇的云服務器通信串口故障數據識別方法。由被動分簇算法確定云服務器通信串口的通信網的漏洞弧段，基于信息熵的量化方法，提取云服務器通信串口通信網漏洞弧段中節點流量數據的熵值特征，將其作為串口故障數據分類方法的分類目標，并以K?means聚類的方式判定云服務器通信串口流量數據的故障類型，實現被動分簇下云服務器通信串口故障數據識別。實驗結果表明，所提方法在多種網絡入侵行為下對云服務器通信串口故障數據識別時，都有較好的識別效果。

關鍵詞： 被動分簇； 云服務器； 通信串口； 故障數據識別； 分類識別； K?means聚類

中圖分類號： TN929.5?34； TP391" " " " " " " " "文獻標識碼： A" " " " " " " " " " " 文章編號： 1004?373X（2025）04?0068?05

Cloud server communication serial port fault data identification under passive clustering

YU Yanpeng， HUI Xianghui

（College of Information and Management Sciences （College of Software）， Henan Agricultural University， Zhengzhou 450000， China）

Abstract： Due to the large and complex data flow in the cloud server communication network， as well as the diversity and dynamic changes in network structure and configuration， traditional active detection or manual analysis methods are difficult to accurately identify vulnerability arcs， which limits the accuracy and efficiency of fault data identification. Therefore， a method of cloud server communication serial port fault data identification based on passive clustering is studied. The passive clustering algorithm is used to determine the vulnerability arc of the communication network of the cloud server communication serial port. Based on the quantification method of information entropy， the entropy characteristics of node traffic data in the vulnerability arc of the cloud server communication serial port communication network are extracted， which are used as the classification target of the serial port fault data classification method. K?means clustering is used to determine the type of fault in the cloud server communication serial port traffic data and realize the identification of cloud server communication serial port fault data under passive clustering. The experimental results show that the proposed method has good recognition performance for cloud server communication serial port fault data under various network intrusion behaviors.

Keywords： passive clustering; cloud server; communication serial port; fault data identification; classification identification; K?means clustering

0" 引" 言

在云服務器中，通信串口是連接服務器與其他設備（如傳感器、外部設備等）的關鍵通道[1?2]。然而，由于各種原因（如硬件老化、軟件錯誤、外部干擾等），通信串口可能會出現故障，導致數據傳輸錯誤、連接丟失等問題，從而影響云服務器的正常運行[3]。傳統的故障識別方法依賴于手動監控和分析，不僅效率低下，而且容易遺漏故障信息。因此，研究通信串口故障數據識別方法具有重要的實際意義和應用價值。

文獻[4]利用集成學習模型提升了通信網絡故障數據的預測準確性和穩定性，但模型參數調優復雜，需要消耗大量實驗和計算資源?！?br>