Security of Separated Data in Cloud Systems with Competing Attack Detection and Data Theft Processes

Area of Science:

• Cloud Computing Security
• Cybersecurity
• Information Assurance

Background:

• Virtualization enables shared physical servers for multiple users, creating vulnerabilities.
• Co-residence attacks exploit shared resources to steal data.
• Data partitioning and early warning agents (EWAs) are proposed security measures.

Purpose of the Study:

• To model and analyze the attack success probability in cloud systems.
• To determine optimal data partitioning and protection policies.
• To minimize user costs while ensuring data security against co-residence attacks.

Main Methods:

• Probabilistic modeling of competing attack detection and data theft processes.
• Analysis of attack success probability based on system parameters.
• Optimization of data partitioning and protection strategies.

Main Results:

• The study quantifies attack success probability considering EWAs and attacker behavior.
• Optimal policies are derived to balance security and cost.
• Parameter sensitivity analysis reveals key factors influencing security.

Conclusions:

• The proposed model provides a framework for securing cloud data against co-residence attacks.
• Effective data partitioning and timely detection are crucial for cloud security.
• The findings guide the development of robust cloud security strategies.