Pattern Match Acceleratorを用いたIoT Edge向けCyber-Securityの研究

カテゴリ: 論文誌(論文単位)

グループ名: 【D】産業応用部門

発行日: 2020/01/01

タイトル(英語): Study on Cyber-Security for IoT Edge Utilizing Pattern Match Accelerator

著者名: 柏山　正守（日本大学大学院理工学研究科），柏山　礼興（東京工業大学工学院機械系），関　弘翔（日本大学理工学部），細野　裕行（日本大学理工学部）

著者名(英語): Masamori Kashiyama (Graduate School of Science and Technology, Nihon University), Reo Kashiyama (Department of Mechanical Engineering, School of Engineering, Tokyo Institute of Technology), Hiroto Seki (College of Science and Technology, Nihon University), Hiroyuki Hosono (College of Science and Technology, Nihon University)

キーワード: サイバーセキュリティ，マルウェア，IoTエッジコンピューティング，パターンマッチアクセラレータ，テクスチャイメージ，高次局所自己相関　　cyber-security，malware，IoT edge computing，pattern match accelerator，texture image，Higher-order Local Auto Correlation (HLAC)

要約(英語): A malware detection algorithm that can be embedded in IoT edge computing is proposed in this study and validated using an emulator. This algorithm, with a pattern match accelerator, reduces the computing cost while maintaining a relatively high detection accuracy. For autonomous driving, complicated IoT edge computing must have a huge amount of embedded program codes. In such a situation, the invasion of malware can lead to compromised cybersecurity. In this study, a pattern match accelerator is implemented for such issues, thereby offering IoT edge computing that detects malware automatically. Edge computing is designed to apply simply structural level analysis algorithms using HLAC mask pattern. We developed a pseudo-emulator system environment and conducted performance confirmation of the proposed technique using 641 chosen samples from six types of malware families. The algorithm's efficiencies demonstrated an identification performance of approximately 80%. In comparison to characteristic extraction using AI, the computing cost was reduced and these processes enable edge computing with high cybersecurity features.

本誌: 電気学会論文誌D（産業応用部門誌） Vol.140 No.1 （2020）

本誌掲載ページ: 15-28 p

原稿種別: 論文／日本語

電子版へのリンク: https://www.jstage.jst.go.jp/article/ieejias/140/1/140_15/_article/-char/ja/