Detection of DDoS Vulnerability in Cloud Computing Using the Perplexed Bayes Classifier

Area of Science:

• Computer Science
• Cybersecurity
• Artificial Intelligence

Background:

• Cloud computing security is paramount due to increasing demand.
• Detecting distributed denial-of-service (DDoS) attacks presents significant challenges due to complex traffic patterns and numerous features.
• Effective defense mechanisms are crucial for widespread cloud adoption.

Purpose of the Study:

• To evaluate the efficacy of a novel perplexed-based classification algorithm for DDoS attack detection in cloud environments.
• To compare the performance of the proposed algorithm against existing machine learning methods and nature-inspired feature selection techniques.

Main Methods:

• A perplexed-based classifier was developed and evaluated with and without feature selection.
• Performance metrics including accuracy, sensitivity, and specificity were used for comparison.
• The proposed classifier was benchmarked against Naïve Bayes, Random Forest, Genetic Algorithm (GA), and Particle Swarm Optimization (PSO).

Main Results:

• The perplexed-based classifier achieved an accuracy of 99%, surpassing existing algorithms.
• Feature selection using the perplexed Bayes classifier demonstrated superior performance compared to GA and PSO, with accuracies 2% and 8% higher, respectively.
• The proposed algorithm proved highly efficient in identifying DDoS attacks within cloud computing systems.

Conclusions:

• The novel perplexed-based classification algorithm offers a highly efficient and accurate solution for DDoS attack detection in cloud computing.
• The proposed feature selection method outperforms nature-inspired algorithms, enhancing detection capabilities.
• This research contributes to robust cloud security by providing an advanced defense mechanism against sophisticated cyber threats.