Artificial intelligence and Blockchain-Assisted Offloading Approach for Data Availability Maximization in Edge Nodes

Gunasekaran Manogaran, Shahid Mumtaz, Constandinos Mavromoustakis, Evangelos Pallis, George Mastorakis

    Research output: Contribution to journalArticlepeer-review

    Abstract

    Mobile Edge Computing (MEC) paradigm is designed to meet the user requirements by providing cloud services at the edge of the user network. Blockchain technology with the EC paradigm is reliable in delivering the edge services depending on user requirements and improving the distributed management of resources at ease. In this article, blockchain-assisted data offloading for Availability Maximization (BDO-AM) is introduced. This proposed approach is presented to thwart the non-probabilistic (NP) hardness problem of data availability due to prolonging backlogs. This approach classifies the different instances of data availability and delivery for the edge-connected end-user services/ applications. The classification is preceded using Nave Bayes classification for identifying the offloading instances to prevent unnecessary backlogs. The probability of data transmission, delivery, and offloading are independently analyzed for their likelihood, and the appropriate available time instances are allocated in a distributed manner. This validation helps to maximize data delivery by reducing the data drops and service delays.

    Original languageEnglish
    JournalIEEE Transactions on Vehicular Technology
    DOIs
    Publication statusAccepted/In press - 2021

    Keywords

    • Blockchain
    • Computational modeling
    • Data Offloading
    • Edge Computing
    • Edge computing
    • Naive Bayes' Classification
    • NP Hardness Problem
    • Reliability
    • Resource management
    • Servers
    • Task analysis

    Fingerprint Dive into the research topics of 'Artificial intelligence and Blockchain-Assisted Offloading Approach for Data Availability Maximization in Edge Nodes'. Together they form a unique fingerprint.

    Cite this