Cloud computing has paved an excellent platform for the emergence of cost-effective technological solutions. However, security and privacy issues still remain as a stringent challenge during service catering. Explicitly, the service utility anomalies are liable to cause severe privacy and security issues in cloud service delivery. So, the overall performance of cloud service consumption and end-user applications’ service levels utility is degraded. The open access and distributed nature of the cloud computing is the major reason for its vulnerability to intruders. The security and privacy in cloud services have many challenges and problems still open for research. This paper proposes an intrusion detection method capable of detecting nine categories of attacks in two stages. This paper focuses on establishing a network-based intrusion detection mechanism using machine learning techniques. A model will be constructed with a supervised learning methodology using historical network flowdata and flowdata collected from the Internet.

云计算为经济高效的技术解决方案的出现铺平了极好的平台。但是，安全和隐私问题仍然是服务提供过程中的严峻挑战。明确地说，服务实用程序异常可能会在云服务交付中引起严重的隐私和安全问题。因此，云服务消耗和最终用户应用程序的服务级别实用程序的整体性能会下降。云计算的开放访问和分布式性质是其易受入侵者攻击的主要原因。云服务中的安全性和隐私性有许多挑战和问题尚待研究。本文提出了一种能够在两个阶段检测到九种攻击的入侵检测方法。本文着重于使用机器学习技术建立基于网络的入侵检测机制。将使用历史网络流量数据和从Internet收集的流量数据，在监督学习方法的基础上构建模型。