Data security is a major concern in cloud computing environments as they provide much scope for intruders to attack. Data centres in cloud environments hold valid information that end-users would conventionally have stored on their computers. Moving information towards centralised services may have an adverse effect on the security of users’ interactions with files kept in cloud cupboard spaces[1], for example accidental or deliberate alterations or deletions of information from the cloud server by the Cloud Service Provider (CSP). This necessitates the deployment of some sort of mechanism to ensure the safety of information integrity[2]. Public sector organisations have much to gain by adopting a cloud computing approach to service delivery in their ICT environments. However, these benefits must be reaped without compromising core requirements and institutional values.
This paper focuses on the security issues that may arise when public sector organisations consider transitioning to an Open Source Software (OSS) Infrastructure as a Service (IaaS) Cloud Infrastructure (OpenStack), although the same issues are likely to be found in other OSS cloud computing software like Apache CloudStack[3], Eucalyptus[4], and OpenNebula[5]. We examine legal implications, regulatory and standards compliance, new attack vectors resulting from vulnerabilities coming from virtualisation technologies, data integrity issues such as encryption and access controls, and security checks to be performed on the services prior to their movement to the cloud. In addition, some of the most important security threats in cloud computing are presented, followed by key recommendations on how to address them, namely security standards and certifications, service provider auditing, secure APIs, transport layer protection, authentication and encryption key management, and cloud service agreements.