.
Popularity of cloud computing have increased especially in the last five years among companies, but also, among individual users. This is due diversity of services provided by cloud, that are organized in software, platforms and infrastructures. Some of the major benefits in adopting cloud are cost reduction, availability of resources, flexibility. Even if it is powerful, cloud computing has some issues, one of them being security. Due to its complex architecture, cloud inherits all security issues of its components. A relatively new approach in cloud security is artificial neural networks, which is used especially in intrusion detection. In this paper we will present some security models based on ANNs applied in cloud computing and which are their results.
As the technology advances fast, the educational and learning environments need to keep up with the trends. Cloud computing is adopted more and more by companies due to its benefits, like availability, cost reduction, on-demand self-service etc. This is also the case of learning environments, because it has been proposed many cloud-based eLearning systems. Still, there are some concerns within such systems, regarding two perspectives: user’s perspective and learning provider’s perspective. A user needs to be sure that its progress or grades or other personal data is accessed by himself or another authorized users (like teachers/instructors). A helpful tool for the user would be a tool that predicts if he/she would pass a final exam, based on past data of user’s behaviour. On the other way, the learning provider needs to be sure that exams are taken by the real user, not by another person on behalf of the real user. In this paper we propose an eLearning system dedicated to mobile devices, such as smartphones or tablets, provided as service, which will focus on analysis of user’s behaviour (based on biometric data, as pulse) in order to determine if he/she is prepared for the course final exam. In this system, the confidentiality of data is assured by homomorphic encryption, in which analysis is done over encrypted data, such that the output will be decrypted only by the user. An additional level of security is represented by authentication method, in which the user’s fingerprint will be used. Also, fingerprint authentication will avoid fraud for situations when the final exam will not be held by the real user. Our proposed system could be used in many areas, like governmental institutions or educational systems.
