https://www.mdu.se/

This paper analyzes the timing performance of a persistent storage designed for distributed container-based architectures in industrial control applications. The timing performance analysis is conducted using an in-house simulator, which mirrors our testbed specifications. The storage ensures data availability and consistency even in presence of faults. The analysis considers four aspects: 1. placement strategy, 2. design options, 3. data size, and 4. evaluation under faulty conditions. Experimental results considering the timing constraints in industrial applications indicate that the storage solution can meet critical deadlines, particularly under specific failure patterns. Comparison results also reveal that, while the method may underperform current centralized solutions in fault-free conditions, it outperforms the centralized solutions in failure scenario. Moreover, the used evaluation method is applicable for assessing other container-based critical applications with timing constraints that require persistent storage.

In this paper, we analyze the failure semantics of a persistent fault-tolerant storage solution for stateful fog applications. This storage system is a container-based solution that provides data availability and consistency in a distributed container-based fog architecture. We evaluate the behavior of this storage system with a formal model that includes all the important time parameters and temporal aspects of the solution. This allows us to verify data consistency and other fault-tolerance properties of our system model while considering application startup latency, together with synchronization intervals and delays. We prove that the solution can tolerate failures at application, node, communication and storage level with the ability to automatically recover from failures and provides data consistency within the synchronization delay defined as t time units, which we can calculate for a given system configuration.

Container-based architectures are widely used for cloud computing and can have an important role in the implementation of fog computing infrastructures. However, there are some crucial dependability aspects that must be addressed to make containerization suitable for critical fog applications, e.g., in automation and robotics. This paper discusses challenges in applying containerization at the fog layer and focuses on one of those challenges: provision of fault-tolerant permanent storage. The paper also presents a container-based fog architecture utilizing so-called storage containers, which combine built-in fault-tolerance mechanisms of containers with a distributed consensus protocol to achieve data consistency.

In our previous work we proposed a fault-tolerant persistent storage for container-based fog architecture. We leveraged the use of containerization to provide storage as a containerized application working along with other containers. As a fault-tolerance mechanism we introduced a replicated data structure and to solve consistency issue between the replicas distributed in the cluster of nodes, we used the RAFT consensus protocol. In this paper, we verify our proposed solution using the UPPAAL model checker. We explain how our solution is modeled in UPPAAL and present a formal verification of key properties related to persistent storage and data consistency between nodes.

Fog computing aims to support novel real-time applications by extending cloud resources to the network edge. This technology is highly heterogeneous and comprises a wide variety of devices interconnected through the so-called fog layer. Compared to traditional cloud infrastructure, fog presents more varied reliability challenges, due to its constrained resources and mobility of nodes. This paper summarizes current research efforts on fault tolerance and dependability in fog computing and identifies less investigated open problems, which constitute interesting research directions to make fogs more dependable. 

Currently, Simulink models can be verified rigorously against design errors or statistical properties. In this paper, we show how Simulink models can be formally analyzed against invariance properties using bounded model checking reduced to satisfiability modulo theories solving. In its basic form, the technique provides means for verification of an underlying model over bounded traces rigorously, however, in general the procedure is incomplete. We identify common Simulink block types and compositions by analyzing selected industrial models, and we show that for some of them the set of non-repeating states (reachability diameter) can be visited with a finite set of paths of finite length, yielding the verification complete. We complement our approach with a tool, called SyMC that automates the following: i) calculation of the reachability diameter size for some of the designs, ii) generation of finite (bounded) paths of the underlying Simulink model and their encoding into SMT-LIB format and iii) checking invariance properties using the Z3 SMT solver. To show the applicability of our approach, we apply it on a prototype implementation of an industrial Simulink model, namely Brake by Wire from Volvo Group Trucks Technology, Sweden. 

Fog computing has been recently introduced to bridge the gap between cloud resources and the network edge. Fog enables low latency and location awareness, which is considered instrumental for the realization of IoT, but also faces reliability and dependability issues due to node mobility and resource constraints. This paper focuses on the latter, and surveys the state of the art concerning dependability and fog computing, by means of a systematic literature review. Our findings show the growing interest in the topic but the relative immaturity of the technology, without any leading research group. Two problems have attracted special interest: guaranteeing reliable data storage/collection in systems with unreliable and untrusted nodes, and guaranteeing efficient task allocation in the presence of varying computing load. Redundancy-based techniques, both static and dynamic, dominate the architectures of such systems. Reliability, availability and QoS are the most important dependability requirements for fog, whereas aspects such as safety and security, and their important interplay, have not been investigated in depth.

Future cyber–physical systems may extend over broad geographical areas, like cities or regions, thus, requiring the deployment of large real-time networks. A strategy to guarantee predictable communication over such networks is to synthesize an offline time-triggered communication schedule. However, this synthesis problem is computationally hard (NP-complete), and existing approaches do not scale satisfactorily to the required network sizes. This article presents a segmented offline synthesis method which substantially reduces this limitation, being able to generate time-triggered schedules for large hybrid (wired and wireless) networks. We also present a series of algorithms and optimizations that increase the performance and compactness of the obtained schedules while solving some of the problems inherent to segmented approaches. We evaluate our approach on a set of realistic large-size multi-hop networks, significantly larger than those considered in the existing literature. The results show that our segmentation reduces the synthesis time by up to two orders of magnitude.

Adaptive requirements for networks with strict timing restrictions do challenge the static nature of the time-triggered communication paradigm. Continuous changes in the network topology during operation require frequent rescheduling, followed by schedule distribution, a process that is excessively time-consuming as it was intended to be performed only during the design phase. The fully-distributed Self-Healing Protocol introduced a collaborative method to quickly modify the local schedules of the nodes during runtime, after link failures. This protocol gets the network back to correct operation in milliseconds, but it assumes that only the nodes are able to modify their local schedules, which limited the achieved improvement. This paper proposes to shift to a semi-distributed strategy, where high-performance nodes are responsible for the nodes and links within a small network segment. These nodes rely on their privileged view of the system in order to reduce the response time, increase the healing success rate, and extend the fault model to include switch failures. 

Industry relies predominantly on manual peer-review techniques for assessing the correctness of system specifications. However, with the ever-increasing size, complexity and intricacy of specifications, it becomes difficult to assure their correctness with respect to certain criteria such as consistency. To address this challenge, a technique called sanity checking has been proposed. The goal of the technique is to assess the quality of the system specification in a systematic and rigorous manner with respect to a formally-defined criterion. Predominantly, the sanity checking criteria, such as for instance consistency, are encoded as reachability or liveness properties which can then be verified via model checking. Recently, a complementary approach for checking the consistency of a system's specification by reducing it to a satisfiability problem that can be analyzed using Satisfiability Modulo Theories has been proposed. In this paper, we compare the two approaches for consistency analysis, by applying them on a relevant industrial use case, using the same definition for consistency and the same set of requirements. Since the bottlenecks of analyzing large systems formally are most often the construction of the model and the time needed to return a verdict, we carry out the comparison with respect to the: i) required effort for generating the analysis model and the latter's complexity, and ii) consistency analysis time. Assuming checking only invariance properties, our results show no significant difference in analysis time between the two approaches when applied on the same system specification under the same definition of consistency. As expected, the main difference between the two comes from the required time and effort of creating the analysis models.