Abstract
A data breach can seriously impact organizational intellectual property, resources, time, and product value. The risk of system intrusion is augmented by the intrinsic openness of commonly utilized technologies like TCP/IP protocols. As TCP relies on IP addresses, an attacker may easily trace the IP address of the organization. Given that many organizations run the risk of data breach and cyber-attacks at a certain point, a repeatable and well-developed incident response framework is critical to shield them. Enterprise cloud possesses the challenges of security, lack of transparency, trust and loss of controls. Technology eases quickens the processing of information but holds numerous risks including hacking and confidentiality problems. The risk increases when the organization outsources the cloud storage services through the vendor and suffers from security breaches and need to create security systems to prevent data networks from being compromised. The business model also leads to insecurity issues which derail its popularity. An attack mitigation system is the best solution to protect online services from emerging cyber-attacks. This research focuses on cloud computing security, cyber threats, machine learning-based attack detection, and mitigation system. The proposed SDN-based multilayer machine learning-based self-defense system effectively detects and mitigates the cyber-attack and protects cloud-based enterprise solutions. The results show the accuracy of the proposed machine learning techniques and the effectiveness of attack detection and the mitigation system.