https://www.mdu.se/

Clouds are powerful computer centers that provide computing and storage facilities that can be remotely accessed. The flexibility and cost-efficiency offered by clouds have made them very popular for business and web applications. The use of clouds is now being extended to safety-critical applications such as factories. However, cloud services do not provide time predictability which creates a hassle for such time-sensitive applications. Moreover, delays in the data communication between clouds and the devices the clouds control are unpredictable. Therefore, to increase predictability an intermediate layer between devices and the cloud is introduced. This layer, the Fog layer, aims to provide computational resources closer to the edge of the network. However, the fog computing paradigm relies on resource-constrained nodes, creating new potential challenges in resource management, scalability, and reliability. Solutions such as lightweight virtualization technologies can be leveraged for solving the dichotomy between performance and reliability in fog computing. In this context, container-based virtualization is a key technology providing lightweight virtualization for cloud computing that can be applied in fog computing as well. Such container-based technologies provide fault tolerance mechanisms that improve the reliability and availability of application execution.  By the study of a robotic use-case, we have realized that persistent data storage for stateful applications at the fog layer is particularly important. In addition, we identified the need to enhance the current container orchestration solution to fit fog applications executing in container-based architectures. In this thesis, we identify open challenges in achieving dependable fog platforms. Among these, we focus particularly on scalable, lightweight virtualization, auto-recovery, and re-integration solutions after failures in fog applications and nodes. We implement a testbed to deploy our use-case on a container-based fog platform and investigate the fulfillment of key dependability requirements. We enhance the architecture and identify the lack of persistent storage for stateful applications as an important impediment for the execution of control applications. We propose a solution for persistent fault-tolerant storage at the fog layer, which dissociates storage from applications to reduce application load and separates the concern of distributed storage. Our solution includes a replicated data structure supported by a consensus protocol that ensures distributed data consistency and fault tolerance in case of node failures. Finally, we use the UPPAAL verification tool to model and verify the fault tolerance and consistency of our solution.

Clouds are large computer centers that offer remote access to computing and storage resources, making them popular for business and web applications. They are now being considered for use in safety-critical applications such as factories, but lack sufficient time predictability, which makes it challenging to use them in these time-sensitive applications. To overcome this limitation, an intermediate layer, the fog layer, is introduced to provide computational resources closer to the network edge. However, this new computing paradigm faces its own challenges in resource management, scalability, and reliability due to resource constrained nodes. Lightweight virtualization technologies like containerization can solve the performance-reliability dichotomy in fog computing and provide built-in fault tolerance mechanisms. By studying a robotic use-case, we realized the critical importance of persistent data storage for stateful applications, such as many control applications. However, container-based solutions lack fault-tolerant persistent storage. In this thesis, we identify new challenges associated with leveraging container-based architectures, particularly the importance of persistent storage for stateful applications. We investigate the design possibilities for persistent fault-tolerant storage and propose a solution adapted to container-based fog architectures and tailored for stateful applications. The solution provides scalability, auto recovery, and re-integration after failures at application and node levels. Key elements are a replicated data structure and a storage container, using a consensus protocol for distributed data consistency and fault tolerance in case of node failures. The fault tolerance and consistency of the solution are modeled and verified, and its timing requirements evaluated. We use simulation to evaluate the timing performance of our solution in larger set-ups. The results of our study show that although adding a consistency protocol introduces a timing overhead, the solution still meets timing requirements for the studied use-case even in presence of a set of relevant faults. By leveraging a four-dimensional approach, we also conduct a comparative analysis of our solution with other approaches from various perspectives, indicating that our solution can be applied in a broader context than initially intended.