QoS-Aware Fault-Tolerant Scheduling for Real-Time Tasks on Heterogeneous Clusters

Abstract

Abstract-Fault-tolerant scheduling plays a significant role in improving system reliability of clusters. Although extensive fault-tolerant scheduling algorithms have been proposed for real-time tasks in parallel and distributed systems, quality of service (QoS) requirements of tasks have not been taken into account. This paper presents a fault-tolerant scheduling algorithm called QAFT that can tolerate one node's permanent failures at one time instant for real-time tasks with QoS needs on heterogeneous clusters. In order to improve system flexibility, reliability, schedulability, and resource utilization, QAFT strives to either advance the start time of primary copies and delay the start time of backup copies in order to help backup copies adopt the passive execution scheme, or to decrease the simultaneous execution time of the primary and backup copies of a task as much as possible to improve resource utilization. QAFT is capable of adaptively adjusting the QoS levels of tasks and the execution schemes of backup copies to attain high system flexibility. Furthermore, we employ the overlapping technology of backup copies. The latest start time of backup copies and their constraints are analyzed and discussed. We conduct extensive experiments to compare our QAFT with two existing schemes-NOQAFT and DYFARS. Experimental results show that QAFT significantly improves the scheduling quality of NOQAFT and DYFARS.

Fault Tolerance in Real-Time Cloud Computing

“…Some works support fault tolerance in cloud environments by scheduling backup copies of failed tasks. For example, QAFT [32] schedules primary and backup copies of tasks on different cloud nodes, considering different QoS levels for the tasks and different speeds for the cloud nodes. However, it only considers independent aperiodic real-time tasks (and not parallel applications), and it does not consider a microservice model as FTRTC does.…”

The impact of checkpointing interval selection on the scheduling performance of real‐time fine‐grained parallel applications in SaaS clouds under various failure probabilities

“…A dedicated VM acts as a central scheduler, responsible for scheduling the component tasks of the application jobs to the other VMs of the cloud. 28 The SaaS cloud under study is shown in Figure 1.…”

A multi-objective co-evolutionary algorithm for energy-efficient scheduling on a green data center

“…He et al [19] developed a rolling-horizon scheduling strategy for the energy constrained distributed real-time embedded systems. Zhu et al [62] presented a fault-tolerant scheduling algorithm called QAFT for real-time tasks with QoS needs on heterogeneous clusters. Zhu et al [63] proposed an adaptive energy-efficient scheduling, AEES, for aperiodic and independent real-time tasks on heterogeneous clusters with dynamic voltage scaling.…”

“…The energy consumption by the processor of computing node can be approximated as α Δ fv t 2 [61][62][63], where f is the processor clock frequency, v is the supply voltage, Δt is the execution time, and α is a constant. The frequency is positive correlated with the voltage, and reducing voltage can lead to low frequency.…”