Heuristic Task Workflow Scheduling in Cloud Using Spot and On-Demand Instances

Suguna, M.; Prakash, Divya; Thangam, D. Yamuna; Shobana, G.

doi:10.1166/jctn.2018.7513

Cited by 5 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To enhance fault tolerance and increase the reliability of virtualized resources, several approaches adopt a mixture of on-demand and Spot instances on scheduling. For instance, the authors of [14], [15], [16] and [17] study the workflow scheduling problem while the authors of [18], [19], [20] and [21] focus on the bag of tasks scheduling. The concept of Latest Time On-Demand (LTO) is introduced in [14] as the latest time in which on-demand instances must be used to guarantee that a workflow's deadline constraint is not violated.…”

Section: Related Workmentioning

confidence: 99%

“…If the difference between LTO and current time is greater than zero (positive slack), tasks are mapped into Spot instances ensuring that the deadline is met while the cost is minimized. The concept of the LTO inspires other works like [15], [16] and [20]. In [14], checkpointing is also applied in three different granularities i.e., every five, ten and thirty minutes.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Proactive Mitigation of the Revocation Risk for Preemptible Virtualized Resources

Oikonomou¹,

Kolomvatsos²,

Anagnostopoulos³

2021

Preprint

View full text Add to dashboard Cite

<div>The provision of resources at the Cloud follows two generic models. The first model guarantees the provided resources for the requested time while the second involves unreliable resources with lower price compared to the former scheme but with no guarantees concerning an unexpected revocation due to a high demand. In this paper, we focus on the latter model and propose a scheme that monitors the course of execution of tasks placed at unreliable resources and decides when to store their current progress avoid jeopardizing intermediate outcomes in unexpected revocations. We rely on the principles of Optimal Stopping Theory (OST) to manage multiple tasks and decide for which task and when we have to save its current status. The outcome is a novel checkpointing mechanism fully aligned with the needs of the dynamics of an unreliable environment. The proposed model builds upon the heterogeneity of the available services in the Cloud and concludes a proactive mitigation approach of the revocation risk for unreliable virtualized resources. We present the theoretical basis of our mechanism and describe the solution of the identified problem. The pros and cons of our approach are evaluated through extensive simulations and a set of performance metrics.</div>

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Proactive Mitigation of the Revocation Risk for Preemptible Virtualized Resources

Oikonomou¹,

Kolomvatsos²,

Anagnostopoulos³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The majority of resource provision techniques are guided by QoS constraints defined by users. There are two basic constraints, i.e., the deadline [26], [24], [22], [21], [18], [16] and the monetary budget [25], [23], [15], [14]. While the deadline constraint is usually satisfied, unreliable resources are used to the maximum extent to limit the monetary costs.…”

Section: Related Workmentioning

confidence: 99%

A Fuzzy Logic Controller for Tasks Scheduling Using Unreliable Cloud Resources

Oikonomou,

Kolomvatsos,

Tziritas

et al. 2020

Preprint

View full text Add to dashboard Cite

The Cloud infrastructure offers to end users a broad set of heterogenous computational resources using the pay-asyou-go model. These virtualized resources can be provisioned using different pricing models like the unreliable model where resources are provided at a fraction of the cost but with no guarantee for an uninterrupted processing. However, the enormous gamut of opportunities comes with a great caveat as resource management and scheduling decisions are increasingly complicated. Moreover, the presented uncertainty in optimally selecting resources has also a negatively impact on the quality of solutions delivered by scheduling algorithms. In this paper, we present a dynamic scheduling algorithm (i.e., the Uncertainty-Driven Scheduling -UDS algorithm) for the management of scientific workflows in Cloud. Our model minimizes both the makespan and the monetary cost by dynamically selecting reliable or unreliable virtualized resources. For covering the uncertainty in decision making, we adopt a Fuzzy Logic Controller (FLC) to derive the pricing model of the resources that will host every task. We evaluate the performance of the proposed algorithm using real workflow applications being tested under the assumption of different probabilities regarding the revocation of unreliable resources. Numerical results depict the performance of the proposed approach and a comparative assessment reveals the position of the paper in the relevant literature.

show abstract