DC 欄位 值 語言 DC.contributor 資訊工程學系 zh_TW DC.creator 戴鴻慈 zh_TW DC.creator Hung-Tsz Tai en_US dc.date.accessioned 2022-7-14T07:39:07Z dc.date.available 2022-7-14T07:39:07Z dc.date.issued 2022 dc.identifier.uri http://ir.lib.ncu.edu.tw:444/thesis/view_etd.asp?URN=109522053 dc.contributor.department 資訊工程學系 zh_TW DC.description 國立中央大學 zh_TW DC.description National Central University en_US dc.description.abstract 後門是一種用於未經授權訪問系統的惡意軟體。攻擊者可以透過後門竊取機密資訊、進行犯罪間諜活動或執行高級持續性威脅攻擊。由於復雜的攻擊需要時間來設置和執行,因此後門是攻擊者重新連接到受害者服務器的一種選擇。雖然很多設備會設置防火牆來防止這類型的攻擊,但後門的多樣性仍然是一個嚴重的問題。一些研究人員甚至發現有後門可以繞過防火牆進入到系統內。 本研究提出了一個機制 BlackCrab,此機制是基於抓取連線 IP 並將其與日誌檔案進行比較的後門檢測方法。藉由在機器上設置不同的服務來獲取其日誌檔案,每種服務會有各自的日誌檔案用來記錄其服務相關的連線或操作事件,當連線 IP 與任一日誌檔案的紀錄相符,便判定此連接為合法的連線。換句話說,如果匹配無效,我們將懷疑此連線是透過後門進入的。除此之外,我們還提供了另一種方法,針對沒有即時記錄日誌的服務進行雜湊值(hash value)比較。 在實驗中,我們選擇了 7 個常見的服務,並從 GitHub 網站和真實世界的後門攻擊中收集了 21 個 Linux 平台上的後門,將其設置在本研究測試的機器上。在 14,737 個合法連接中,通過檢查的準確率達到 99.6%,並且在 21個後門中,BlackCrab 可以檢測到其中的 19 個。經過測試,此機制對原始系統的開銷可以忽略不計。 zh_TW dc.description.abstract Backdoor is a type of malware that is used to get unauthorized access to a system. When attackers enter through the device successfully, they can steal sensitive information, do criminal-espionage activity, or execute an advanced persistent threat attack. Because sophisticated attack takes time to set up and execute, backdoor is a choice for criminals to reconnect to the victim server. Although many devices have firewalls to prevent attacks, the variety of backdoors is still a serious problem. Researchers even found that some backdoors can bypass the firewall. We propose a system, BlackCrab, based on the method of capturing the connection IP and comparing it with log files to detect backdoors. We compare the log files of the services we have set up on the machine to determine whether the connection is legal. Besides, we provide another method, hash value comparison, for some services that do not have real-time log records. In our implementation, we chose seven common services and collected 21 backdoors targeting the Linux platform from the GitHub project and real-world backdoor samples. With 14,737 legitimate connections, 99.6% accuracy of passing inspections is achieved. Of 21 backdoors, 19 of them can be detected by BlackCrab. The overhead introduced by our system is negligible. en_US DC.subject 後門 zh_TW DC.subject 日誌檔案 zh_TW DC.subject 即時檢測 zh_TW DC.subject Backdoor en_US DC.subject log en_US DC.subject real-time detection en_US DC.title BlackCrab: A Connection-Based Real-time Backdoor Detection Mechanism en_US dc.language.iso en_US en_US DC.type 博碩士論文 zh_TW DC.type thesis en_US DC.publisher National Central University en_US