Enhancing Cloud Security in DDoS: A Holistic Approach Leveraging Machine Learning Techniques
Manju C Nair, B. Ankayarkanni · 2024
Industrial application development, distribution, and deployment are impacted by the cloud computing industry's explosive growth. The majority of enterprises appear to be shifting in the last few days toward adopting cloud environments and different cloud-based services. Ensuring a secure and reliable environmental solution throughout cloud space is crucial for protecting and safeguarding the transactions conducted by enterprises. Static verification of cloud user behavior using methods like support vector machines and linear regression have been tried to discover pre-defined dangers and advance cyber- security in the market. These security solutions have limited functionality because they are static. Making decisions about access control entails executing a block or permit action. Additionally, the prior approaches have issues with data security for endpoints that are not cloud-managed. Furthermore, distributed denial-of- service (DDoS) and data privacy assaults are the most prevalent types of Cloud security attacks, with 16% and 14% of users, respectively. In order to address the aforementioned problems, this research focuses on creating a novel security solution for effective DDoS attack detection utilizing machine learning (ML) approaches. The primary goal is to influence the direction of cloud security in the future by utilizing machine learning algorithms, which can offer automated and adaptable methods for improving security in cloud environments. ML can provide solutions that go beyond focusing only on pattern recognition and analysis in sensitive data, to include full algorithms for securing enterprise data across all cloud apps. The more accurate model is trained with 99.6% accuracy by the random forest (RF) classifier, according to experimental results. The suggested machine learning algorithms are put to the test, the outcomes are confirmed, and the algorithm's effectiveness is assessed by contrasting it with other currently used techniques.