Scholarlink

<p>Authorization is a fundamental service in the Hadoop environment. It ensures data and protocol security while granting access exclusively to authorized users, allowing them to perform operations such as data storage, processing, and file transfer in a controlled manner. However, maintaining effective authorization remains challenging due to vulnerabilities in the Hadoop ecosystem and potential abnormal user behavior. Various authorization protocols have been developed based on different assumptions about the Hadoop environment, each aiming to ensure reliable and efficient access control through distinct techniques. While these techniques have their strengths and limitations, authorization can be further enhanced by integrating additional security mechanisms. In this work, we propose a trust-based authorization model to assist the NameNode in making access control decisions, particularly when handling uncertain or untrusted authenticated users. This model addresses two key questions: (1) should the authenticated user be authorized to perform the requested operation? (2) how should the request be executed? Our approach leverages machine learning alongside existing security tools, such as monitoring systems and intrusion detection mechanisms, to identify intrusions or unauthorized operations. By continuously analyzing incoming requests and monitoring the Hadoop environment, our method enhances security while maintaining operational efficiency. Additionally, we integrate elephant herding optimization to enable a distributed and adaptive access control mechanism, optimizing user satisfaction, minimizing security risks, and reducing resource consumption. The proposed model is evaluated against a generic authorization framework based on role-based access control and attribute-based access control using different performance metrics. Experimental results demonstrate the effectiveness of our approach in detecting unauthorized access. Furthermore, the machine learning-based intrusion detection method is assessed using widely recognized datasets, considering key metrics such as accuracy, recall, precision, F1-score, specificity, AUC-ROC, and execution time. The results illustrate that the applied ML techniques are effective and can achieve optimal performance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A trust-driven optimization model for reliable authorization in Hadoop Environment

  • 期刊论文

  • 发表日期：

    2025年4月28日

查看全文
  • Nadia Battat,
  • Abdallah Makhoul

摘要

Authorization is a fundamental service in the Hadoop environment. It ensures data and protocol security while granting access exclusively to authorized users, allowing them to perform operations such as data storage, processing, and file transfer in a controlled manner. However, maintaining effective authorization remains challenging due to vulnerabilities in the Hadoop ecosystem and potential abnormal user behavior. Various authorization protocols have been developed based on different assumptions about the Hadoop environment, each aiming to ensure reliable and efficient access control through distinct techniques. While these techniques have their strengths and limitations, authorization can be further enhanced by integrating additional security mechanisms. In this work, we propose a trust-based authorization model to assist the NameNode in making access control decisions, particularly when handling uncertain or untrusted authenticated users. This model addresses two key questions: (1) should the authenticated user be authorized to perform the requested operation? (2) how should the request be executed? Our approach leverages machine learning alongside existing security tools, such as monitoring systems and intrusion detection mechanisms, to identify intrusions or unauthorized operations. By continuously analyzing incoming requests and monitoring the Hadoop environment, our method enhances security while maintaining operational efficiency. Additionally, we integrate elephant herding optimization to enable a distributed and adaptive access control mechanism, optimizing user satisfaction, minimizing security risks, and reducing resource consumption. The proposed model is evaluated against a generic authorization framework based on role-based access control and attribute-based access control using different performance metrics. Experimental results demonstrate the effectiveness of our approach in detecting unauthorized access. Furthermore, the machine learning-based intrusion detection method is assessed using widely recognized datasets, considering key metrics such as accuracy, recall, precision, F1-score, specificity, AUC-ROC, and execution time. The results illustrate that the applied ML techniques are effective and can achieve optimal performance.