Information Technology Reference
In-Depth Information
Temporal Verification for Scientific Cloud Workflows:
State-of-the-Art and Research Challenges
Qiudan Wang
1
, Xiao Liu
1
, Zhou Zhao
1
, and Futian Wang
2
1
Shanghai Key Laboratory of Trustworthy Computing, East China Normal University, China
2
School of Computer Science and Technology, Anhui University, China
qiudan.wang@outlook.com, xliu@sei.ecnu.edu.cn,
zzhao901@gmail.com, wft@ahu.edu.cn
Abstract.
Cloud computing is establishing itself as the latest computing para-
digm in recent years. As doing science in the cloud is becoming a reality, scien-
tists are now able to access public cloud centers and employ high-performance
computing resources to run scientific applications. However, due to the dynam-
ic nature of the cloud environment, the usability of scientific cloud workflow
systems can be significantly deteriorated if without effective service quality as-
surance strategies. Specifically, workflow temporal verification as the major
approach for workflow temporal QoS (Quality of Service) assurance plays a
critical role in the on-time completion of large-scale scientific workflows. Great
efforts have been dedicated to the area of workflow temporal verification in re-
cent years and it is high time that we should define the key research issues for
scientific cloud workflows in order to keep our research on the right track. In
this paper, we systematically investigate this problem and present four key re-
search issues based on the introduction of a generic temporal verification
framework. Meanwhile, state-of-the-art solutions for each research issue and
open challenges are also presented. Finally, SwinDeW-V, an ongoing research
project on temporal verification as part of our SwinDeW-C cloud workflow
system, is also demonstrated.
Keywords:
Scientific Workflow, Workflow Temporal Verification, Cloud
Computing, Quality of Service.
1
Introduction
Many complex e-science applications such as climate modeling, earthquake model-
ing, weather forecast, Astrophysics and high energy physics require high-performance
computing infrastructures [15]. In addition, since scientific applications are often
collaborative processes among groups of scientists and require geographically distri-
buted scientific equipment and data resources, modern scientific applications can
normally be carried out in the form of scientific workflows [55]. A survey of scientif-
ic workflow systems can be found at [54]. The last decade has seen a significant de-
velopment of distributed workflow systems with the grid computing paradigm. The
workflow enactment service of grid workflow management systems may be built on
