Many businesses are moving to the cloud to increase efficiency, flexibility, fault tolerance, collaboration, and competitiveness. These benefits are achieved by increasing heavily resource sharing. Many e-commerce sites share the same physical server, the same OS image, the same memory pages. But increased sharing usually means potential information leakage. Security in cloud computing environments is a big challenge.
It is time to increase the intelligence of nodes and network devices. By observing the behaviour of the different services in a data center, the network devices can dynamically develop a model of normal operation. Any attack or fault can be detected as a deviation from the normal operation model. The aims of this project are therefore:
- Study, design, analyze, and optimize different risk classifiers using unsupervised learning techniques.
- Design and analyze distributed actuation policies to confine a potential attack.
- Design, analyze, and optimize a distributed reputation system to combine different classifiers and actuators.
- Develop HW versions of the algorithms to integrate them in real state-of-the-art network devices based on the NetFPGA SUME platform.
During the development of this project, you will work in close cooperation with different companies. If successful, this work will have have a huge impact in current cloud infrastructures.
Applicants should have, or expect to obtain by the start date, a degree in a relevant engineering or science subject.
A doctoral training grant from the GREENSTACK Project will be available during 2016 for this work, and we expect this project will be fully funded by mid 2016 for 3 additional years. Applications for this fully funded research studentship is invited from suitably qualified EU students.
Further information is available from the project supervisor, José M. Moya.
Applications should be made on-line through this job board.
Our team values diversity and is committed to equality of opportunity.
|Job Category||I3 - Investigador No Doctor|