Replication is a key mechanisms for building systems that tolerate faults and malicious intrusions. Our research focusses on two main areas:
- How can you design more efficient replication systems using virtualization technology? How can you support proactive and reactive recovery without sacrificing availability and performance? How can you build Byzantine fault tolerant systems with low resource requirements?
- How can you design efficient total-order broadcast algorithm? How can you tailor such algorithm to the specifics of the operating environment? How can you make these systems self-configurung and self-optimizing?
Detecting and analyzing problems is a challenge especially of systems span multiple administrative domains. As a simple example, a cloud client has very little access possibilities to the cloud provider's infrastructure.
- What methods, mechanisms and interfaces facilitate the detection and analysis of problems in multi-domain environments such as public clouds?
- How can you detect, analyze, understand in detail complex multi-stage attacks?
Replication infrastructures and diagnosis system can both benefit from suitable functionality at the operating system and hypervisor level.
- What mechanisms and interfaces can be implemented at the operating system / hypervisor level?
- How to enhance the protection of integrity and confidentiality of data and services in the cloud using trusted computing techniques?
- How to protect the OS and hypervisor themselves against attacks?