Scribd Logo
Upload

Welcome to Scribd!

  • IoT Malware Detection Pavlovsky

    Uploaded by

    kgb
    15 views6 pages
    This document discusses addressing resource optimization challenges for IoT malware detection. It proposes using a one-time classifier that only needs to update when internet connectivity is available to overcome connectivity issues. It also suggests creating a lightweight virtual environment called a Security Interpreter to address the heterogeneous OS and hardware fragmentation issue. The document concludes that this approach could result in an affordable and robust malware detection product for IoT devices within two years, and further work is needed to determine if one overall model or multiple models per malware family is sufficient.

    Original Description:

    iot

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    This document discusses addressing resource optimization challenges for IoT malware detection. It proposes using a one-time classifier that only needs to update when internet connectivity is available to overcome connectivity issues. It also suggests creating a lightweight virtual environment called a Security Interpreter to address the heterogeneous OS and hardware fragmentation issue. The document concludes that this approach could result in an affordable and robust malware detection product for IoT devices within two years, and further work is needed to determine if one overall model or multiple models per malware family is sufficient.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Flag for inappropriate content
    15 views6 pages

    IoT Malware Detection Pavlovsky

    Uploaded by

    kgb
    This document discusses addressing resource optimization challenges for IoT malware detection. It proposes using a one-time classifier that only needs to update when internet connectivity is available to overcome connectivity issues. It also suggests creating a lightweight virtual environment called a Security Interpreter to address the heterogeneous OS and hardware fragmentation issue. The document concludes that this approach could result in an affordable and robust malware detection product for IoT devices within two years, and further work is needed to determine if one overall model or multiple models per malware family is sufficient.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Flag for inappropriate content
    of 6

    IoT Malware Detection Addressing

    Resource Optimization
    N AT H A N PAV L O V S K Y:
    PAV L O V S N @ E M A I L . S C . E D U
    CONOR BABIN: CBABIN@EMAIL.SC.EDU
    THAHIMUM HASSAN: THASSAN@EMAIL.SC.EDU
    B R A N D O N RY D E R : B RY D E R @ E M A I L . S C . E D U
    S T E V E N E D WA R D S : S T E V E N G E @ E M A I L . S C . E D U
    LUKE IMHOLZ: LIMHOLZ@EMAIL.SC.EDU
    Addressing the problem

    IoT devices have limited


    memory and power, making
    balancing malware detection
    algorithms’ efficiency,
    accuracy, and resource usage
    a significant issue.
    Challenging Aspect: Internet Connectivity
    Overcoming Connectivity One Time Classifier
    Issues Resolved by using a one-time
    Malware detection systems classifier that only needs to
    that require internet update when server connectivity
    connectivity faces the issue is available
    of failure without internet. • We propose a CNN neural
    This impedes the usage of a network classifier; is trained
    centralized cloud-based on binaries that are converted
    algorithm for continuous to gray-scale images
    detection process.
    Challenging Aspect: Heterogeneous OSes
    Problem: large fragmentation Security Interpreter (SECI)
    of the IoT market among Creates a lightweight virtual
    different operating systems environment on machines
    and hardware configurations • Limited architecture
    • Little memory & power used
    • For most important OSes
    Challenging Aspect:
    Limited Memory and Processing Power
    Main training of classifier is
    Limited computational
    offloaded to the server.
    resources restricts IoT
    devices from using resources
    Harnessing GPU processing
    for the purpose of malware
    power and parallelization
    detection.
    techniques will help power
    classifier execution on end hosts
    Conclusion and Future Work
    • It is estimated that this approach can result in an affordable,
    robust, scalable, and commercially-viable product for market
    deployment in two years.

    • Further work can be done to determine if it is sufficient to train


    one overall malware-detection model or if multiple models for
    different malware families need to be trained.
    • GOAL: maintain at least 95% detection accuracy

    You might also like