A Scheduling Method for Tasks and Services in IIoT Multi-Cloud Environments

Owing to more resources and higher scalability, cloud environments are more prevalent than traditional local servers to deploy Industrial Internet of Things (IIoT) applications. In a multi-cloud environment, there are multiple alternative clouds to deploy applications. These clouds have different resources and network conditions, and at the same time, also the IIoT applications have diverse requirements and priorities. Thus, in order to allocate the necessary resources to applications, there is a need for scheduling algorithms that manage the process of cloud selection and resource allocation. Besides, in practical scenarios, two types of applications can be identified, namely, services and tasks, with the former being long-lasting executables and the latter being code running for a short amount of time. Even though these two types of applications should be able to be deployed at the same time, to the best of our knowledge, the existing algorithms can only schedule one type of the two. In this paper, we propose an algorithm named Multi-Cloud Application Scheduling Genetic Algorithm (MCASGA), which can schedule both services and tasks at the same time. MCASGA can make scheduling schemes according to application priorities, application dependence, network latency, network bandwidth, CPU, memory, and storage. Our simulated experiments show that in scenarios with different service-to-task ratios, MCASGA outperforms four existing algorithms in the aspects of application acceptance rate, task completion time, and application makespan. The results show that when MCASGA completes 90% of the tasks, other algorithms can only complete less than about 50%.