Railway applications such as train dispatching and traffic management are safety sensitive, therefore they require high availability and reliability. However the railway applications nowadays have a localized and centralized service paradigm, leading to safety concerns: The service maintenance responsiblity highly depends on human. The consequence is that only a low level fault tolerance can be achieved, including inefficient failover and quality of service degradation when failure occurs. This can be addressed by a cloud-based solution, named RailCloud, that aims to 1) shift the applications from local machines to a cloud hosted in a data center, and 2) reinforce the reliability guarantee of the cloud.
Cloud computing provides a new form of resource provisioning, where services and applications are no longer hosted on local computing resources, but instead on the shared resources that are provided by remote infrastructures. Besides the most common form of cloud offering that delivering virtual machines to the customers, Platform as a Service (PaaS) clouds are getting increasingly popular among software developers. PaaS Clouds aim to help customers to quickly deploy and run their applications without considering infrastructure management tasks, and therefore considered as good bases for implementing RailCloud.
High availability and reliability of applications can be achieved by using Byzantine Fault Tolerance (BFT) protocols to build state machine replicas. BFT services can tolerate not only crash-stop but also arbitrary failures of replicas. To fulfill the reliability requirement of railway applications, BFT protocols need to be integrated into the PaaS cloud. Integration of BFT into the PaaS cloud offers BFT to the customers as a built-in service, so that they can simply choose such service when deploying railway applications, and the replicas would be automatically generated and coordinated. This way, railway applications are guaranteed to tolerate Byzantine failures, without complicated strategy and operations at the customer side.
| Title | Type | Supervisor | Status |
|---|---|---|---|
| Automated Deployment and Evaluation of BFT Systems Using a Metal-as-a-Service Cloud | Master Thesis | Dr. Bijun Li | finished |
| Improve Dynamic State Partitioning in Parallel BFT Systems | Bachelor Thesis | Dr. Bijun Li | finished |
| Automated Deployment and Evaluation of Fault Tolerant Systems Using a Metal-as-a-Service Cloud | Master Thesis | Dr. Bijun Li | finished |
| A Multi-leader Approach of Byzantine Fault Tolerance | Master Thesis | Dr. Bijun Li | finished |
| Smart service state partitioning in BFT | Bachelor Thesis | Dr. Bijun Li, Wenbo Xu | finished |
| Dynamic state partitioning in Byzantine Fault Tolerance system | Project Thesis | Dr. Bijun Li, Wenbo Xu | finished |
This research project is funded by Siemens international Rail Automation Graduate School (iRAGS).