The advent of cloud computing has made a global impact by providing on-demand services, elasticity, scalability, and flexibility, hence delivering cost-effective resources to end users in pay-as-you-go manner. However, securing cloud services against vulnerabilities, threats, and modern attacks remains a major concern. Application layer attacks are particularly problematic because they can cause significant damage and are often difficult to detect, as malicious traffic can be indistinguishable from normal traffic flows. Moreover, preventing Distributed Denial of Service (DDoS) attacks is challenging due to its high impact on physical computer resources and network bandwidth. This study examines new variations of DDoS attacks within the broader context of cyber threats and utilizes Artificial Intelligence (AI)-based approaches to detect and prevent such modern attacks. The conducted investigation determines that the current detection methods predominantly employ collectively, hybrid, and single Machine Learning (ML)/Deep Learning (DL) techniques. Further, the analysis of diverse DDoS attacks and their related defensive strategies is vital in safeguarding cloud infrastructure against the detrimental consequences of DDoS attacks. This article offers a comprehensive classification of the various types of cloud DDoS attacks, along with an in-depth analysis of the characterization, detection, prevention, and mitigation strategies employed. The article presents, an in-depth analysis of crucial performance measures used to assess different defence systems and their effectiveness in a cloud computing environment. This article aims to encourage cloud security researchers to devise efficient defence strategies against diverse DDoS attacks. The survey identifies and elucidates the research gaps and obstacles, while also providing an overview of potential future research areas.

中文翻译：

---

全面审查漏洞和 AI 支持的 DDoS 攻击防御，以保护云服务


云计算的出现通过提供按需服务、弹性、可扩展性和灵活性产生了全球影响，从而以即用即付的方式向最终用户提供具有成本效益的资源。然而，保护云服务免受漏洞、威胁和现代攻击仍然是一个主要问题。应用层攻击尤其成问题，因为它们可能造成重大损害，并且通常难以检测，因为恶意流量与正常流量无法区分。此外，由于分布式拒绝服务 (DDoS) 攻击对物理计算机资源和网络带宽的影响很大，因此防止分布式拒绝服务 (DDoS) 攻击具有挑战性。这项研究在更广泛的网络威胁背景下研究了 DDoS 攻击的新变体，并利用基于人工智能 (AI) 的方法来检测和防止此类现代攻击。所进行的调查确定，当前的检测方法主要采用集体、混合和单一机器学习（ML）/深度学习（DL）技术。此外，对各种 DDoS 攻击及其相关防御策略的分析对于保护云基础设施免受 DDoS 攻击的有害后果至关重要。本文对各种类型的云 DDoS 攻击进行了全面分类，并对所采用的特征、检测、预防和缓解策略进行了深入分析。本文深入分析了用于评估不同防御系统及其在云计算环境中的有效性的关键性能指标。本文旨在鼓励云安全研究人员针对各种 DDoS 攻击制定有效的防御策略。 该调查确定并阐明了研究差距和障碍，同时还概述了未来潜在的研究领域。