Emerging Internet of Things (IoT)-enabled cyber-physical applications, such as autonomous critical infrastructure inspection, demand low-latency, energy-efficient, and reliable execution across resource-constrained edge devices with heterogeneous multicore processors and diverse sensing and actuating capabilities, in collaboration with a hub device and a cloud server. These workflow-based applications comprise interdependent tasks that must be executed under stringent deadline, reliability, capability, memory, storage, and energy constraints. Given their critical nature, exact optimization is necessary to obtain optimal schedules that ensure dependable operation. Existing scheduling approaches, both exact and heuristic, fail to jointly address all these objectives and constraints. To this end, we propose an exact multi-objective and multi-constrained workflow scheduling approach for edge-hub-cloud cyber-physical systems, based on continuous-time mixed integer linear programming. The proposed formulation jointly optimizes latency, energy, and reliability, while holistically addressing timing and resource constraints. To enhance reliability while avoiding the overhead of unnecessary task replicas, it selectively employs task duplication. We evaluate our approach against a widely used heuristic, which we extend to ensure a fair and meaningful comparison, using a real-world IoT workflow and synthetic task graphs of varying sizes, across different system configurations and objective trade-offs. The proposed method consistently outperforms the heuristic, achieving up to 29.83%, 33.96%, and 28.49% average improvements in latency, energy, and reliability, respectively, while attaining practical runtimes. Overall, the experimental results demonstrate the effectiveness of our approach under various system configurations and objective trade-offs, and show its practical scalability to task graphs of sizes relevant to the targeted applications and system architecture.

To address the challenges of imbalanced multi-class datasets typically used for rare event detection in critical cyber-physical systems, we propose an optimal, efficient, and adaptable mixed integer programming (MIP) ensemble weighting scheme. Our approach leverages the diverse capabilities of the classifier ensemble on a granular per class basis, while optimizing the weights of classifier-class pairs using elastic net regularization for improved robustness and generalization. Additionally, it seamlessly and optimally selects a predefined number of classifiers from a given set. We evaluate and compare our MIP-based method against six well-established weighting schemes, using representative datasets and suitable metrics, under various ensemble sizes. The experimental results reveal that MIP outperforms all existing approaches, achieving an improvement in balanced accuracy ranging from 0.99% to 7.31%, with an overall average of 4.53% across all datasets and ensemble sizes. Furthermore, it attains an overall average increase of 4.63%, 4.60%, and 4.61% in macro-averaged precision, recall, and F1-score, respectively, while maintaining computational efficiency.

The emerging edge-hub-cloud paradigm has enabled the development of innovative latency-critical cyber-physical applications in the edge-cloud continuum. However, this paradigm poses multiple challenges due to the heterogeneity of the devices at the edge of the network, their limited computational, communication, and energy capacities, as well as their different sensing and actuating capabilities. To address these issues, we propose an optimal scheduling approach to minimize the overall latency of a workflow application in an edge-hub-cloud cyber-physical system. We consider multiple edge devices cooperating with a hub device and a cloud server. All devices feature heterogeneous multicore processors and various sensing, actuating, or other specialized capabilities. We present a comprehensive formulation based on continuous-time mixed integer linear programming, encapsulating multiple constraints often overlooked by existing approaches. We conduct a comparative experimental evaluation between our method and a well-established and effective scheduling heuristic, which we enhanced to consider the constraints of the specific problem. The results reveal that our technique outperforms the heuristic, achieving an average latency improvement of 13.54% in a relevant real-world use case, under varied system configurations. In addition, the results demonstrate the scalability of our method under synthetic workflows of varying sizes, attaining a 33.03% average latency decrease compared to the heuristic.

In this study, we propose and investigate a scheduling technique for linear workflow (LW) jobs with dynamically adjustable exit tasks in a distributed system. The proposed approach takes into account the probability for a LW job to permit partial computations for its exit task, under specific conditions. We investigate via simulation the case where none, half or all of the LW jobs in the workload allow partial computations, under different scenarios of load conditions and service demand variability. The simulation results provide useful insights into how the probability for a LW job to permit partial computations for its exit task, as well as the system load, affect the system performance. The results also demonstrate that partial computations, under the proposed scheduling scheme, can reduce the mean response time of the LW jobs, incurring only an insignificant degradation in their average result precision.

In distributed environments, applications are usually complex and computationally demanding, having a linear workflow (LW) structure. Additionally, such LW jobs may also have different priorities for processing. This entails the danger of long delays for low priority jobs. Furthermore, transient software failures may occur during the execution of the workload. Consequently, resource allocation, scheduling and fault tolerance are three crucial aspects that should be efficiently and effectively addressed in such environments, in order to achieve good system performance. To this end, in this paper we investigate the resource allocation and scheduling of LW jobs that arrive dynamically in an environment of distributed resources. We consider that the LW jobs have different priorities and that transient software failures may occur during their execution. A novel scheduling technique is proposed, which takes into account the ageing priorities of the LW jobs, as well as the resulting scheduling overhead. We examine the performance of three routing strategies in this framework, under various load cases and different failure probabilities, taking also into account their implementation complexity. The simulation results reveal how each routing strategy is affected in each of the examined scenarios.

Summary Cloud bursting is a concept originating from the hybrid cloud computing paradigm. During workload spikes, the local resources of the private cloud are supplemented by resources in the public cloud. This technique could also be applied in a fog computing environment, in order to handle workload fluctuations, by offloading applications to the cloud. Toward this direction, in this article, we propose a strategy for the utilization of supplementary cloud resources, in order to assist in the processing of Internet of Things workflow jobs that arrive dynamically in a fog environment. As the cloud involves higher data transfer latency and monetary cost, our approach takes into account these two factors, in addition to the real‐time constraints of the workload. The proposed scheduling heuristic is based on the tradeoff between performance and monetary cost. During resource selection, different contribution factors of these two parameters are assessed. Furthermore, the proposed scheduling method is compared against a baseline policy that utilizes only the fog resources. The simulation experiments were carried out under different sizes of workflow input data and for workloads with soft and hard deadlines.

Summary Software as a Service (SaaS) cloud computing has emerged as an attractive platform to tackle various problems of the traditional software distribution model, such as the requirement to acquire and maintain expensive hardware and software infrastructure. SaaS, however, involves many challenges, mainly due to the heterogeneity and multitenancy of the underlying host environment, as well as the nature of the applications executed on such platforms. Applications are usually bags‐of‐tasks, consisting of independent component tasks that can be executed in any order, featuring different degrees of variability in their computational demands. Furthermore, according to the service level agreement between the cloud provider and the end‐users, the execution of such applications must typically complete within a deadline, providing results of acceptable quality. Consequently, one of the most important aspects of SaaS cloud computing is the effective scheduling of multiple parallel applications, avoiding any service level agreement violations. Towards this direction, our contribution in this paper is twofold: (1) We enhance some of the most commonly used scheduling algorithms for bag‐of‐tasks applications, by utilizing approximate computations, and (2) we investigate the impact of different levels of variability in the computational demands of the applications on the performance of the examined heuristics.

Distributed resources required for processing complex parallel applications are becoming larger in scale day-by-day. They need to be used efficiently in order to provide Quality of Service (QoS). Therefore, resource allocation and scheduling are of paramount importance. In this paper we investigate issues involved with the scheduling of parallel jobs that are bag-of-task-chains in distributed systems. Two task-chain scheduling techniques in two different cases of resource allocation are studied. The performance of the task-chain scheduling algorithms is evaluated via simulation, under different cases of system workload. The experimental results show that the performance of the task-chain scheduling strategies depends on the employed resource allocation techniques.

Summary As the adoption of Software as a Service (SaaS) cloud computing continues to gain momentum, the arising challenges of scheduling parallel applications on such platforms need to be addressed. Due to the complexity and the fine‐grained parallelism of the workload, as well as the multi‐tenancy of the underlying host environment, end‐user applications are usually prone to transient software failures. Therefore, fault tolerance is one of the most crucial aspects of scheduling in SaaS clouds. It is usually achieved through application‐directed checkpointing. However, selecting an appropriate checkpointing interval is not a trivial task. Unnecessary frequent checkpointing may degrade the system performance. On the other hand, infrequent checkpointing may lead to greater recovery time and thus poorer performance. Consequently, the checkpointing interval must be selected taking into account the failure probability, as well as the nature of the workload. Towards this direction, we investigate via simulation the impact of checkpointing interval selection on the performance of a SaaS cloud, where fine‐grained parallel applications with firm deadlines and approximate computations are scheduled for execution, under various failure probabilities. The simulation results are analyzed, in an attempt to shed light on the relation between the checkpointing interval and failure probability.

As cloud computing continues to gain momentum, big data analytics are now offered as Software as a Service (SaaS). Besides the heterogeneity and multi-tenancy of the underlying virtualized environment, scheduling such real-time, data-intensive, embarrassingly parallel applications in a SaaS cloud involves another serious challenge: data locality. Consequently, data-aware scheduling policies should be employed, in order to effectively exploit data locality, while at the same time taking into account the other attributes of the workload and the characteristics of the resources. Towards this direction, we investigate via simulation the impact of data locality on the performance of a SaaS cloud, where real-time, data-intensive bags-of-tasks are scheduled dynamically, under various data availability conditions. A non-data-aware baseline scheduling policy is compared with two proposed data-aware heuristics, in an attempt to shed light on the effect of data locality awareness on the system performance.