A Novel Hadoop Security Model for Addressing Malicious Collusive Workers

Area of Science:

• Computer Science
• Cloud Computing
• Big Data Analytics

Background:

• Hadoop is a distributed system for big data processing using MapReduce.
• Existing security solutions for Hadoop in cloud environments face challenges like high overhead and delayed detection of malicious workers.

Purpose of the Study:

• To propose a novel model to enhance the security of cloud environments against untrusted workers in Hadoop.
• To develop a component to detect both non-collusive and collusive malicious workers.

Main Methods:

• Development of a Malicious Workers' Trap (MWT) component integrated into the master node.
• Implementation and performance analysis of the MWT model in a Hadoop cluster.

Main Results:

• The proposed MWT model accurately detects malicious workers.
• MWT demonstrates minor processing overhead compared to vanilla MapReduce and Verifiable MapReduce (V-MR).
• The model effectively balances security and usability in Hadoop clusters.

Conclusions:

• The MWT model offers an effective solution for detecting malicious workers in Hadoop cloud environments.
• MWT provides a secure and usable framework for big data processing, addressing limitations of existing approaches.