Information Technology Reference
In-Depth Information
checkpoint selection and temporal verification. The workflow execution state can be
displayed by a watch list which contains runtime information such as time submitted,
time finished, percentage of completion, service status and many other real-time and
possible statistic data. When the QoS violations such as temporal violations are de-
tected, alter messages will be sent to the QoS violation handling component. The QoS
violation handling component which includes the temporal violation handling strate-
gies will analyze the workflow execution state and the QoS requirement specification
to decide further actions such as workflow rescheduling and additional resource re-
cruitment through the interactions with the application provision services.
Currently, SwinDeW-V has provided fully support for high temporal QoS for
scientific workflows [26]. In the future, SwinDeW-V will explore more on the sup-
port for business cloud workflows. Specifically, those open issues proposed in Section
3 need to be further investigated. Our ultimate objective is that SwinDeW-V can be
developed as an independent software component which can be easily adopted by any
cloud workflow systems including both scientific and business cloud workflows to
facilitate the functionalities of the temporal verification framework.
5
Conclusion
Workflow temporal verification is the major approach to guarantee the on-time com-
pletion of workflow applications. In the last few years, a lot of research efforts have
been dedicated to the temporal verification in scientific workflow systems. However,
with the emergence of cloud computing, there are some open challenges for the tem-
poral verification in scientific cloud workflow systems. In this paper, through the
introduction of a generic temporal verification framework, four basic research issues
for scientific cloud workflow temporal verification including temporal constraint
setting, temporal checkpoint selection, temporal verification and temporal violation
handling were first presented with representative and state-of-the-art solutions. These
four basic research issues can lay out a clear research road map for scientific cloud
workflow temporal verification. Afterwards, three open research challenges including
the forecasting strategy for scientific cloud workflow activity durations, the monitor-
ing of many parallel computing tasks, and temporal violation handling for scientific
cloud workflows are discussed with the introduction of some potential research direc-
tions. This paper provided a systematic overview of the current research work and
some open challenges for scientific cloud workflows temporal verification in the fu-
ture. The result of this paper is useful for those researchers who are interested in time
related quality assurance in scientific cloud workflow systems.
Acknowledgement. The research work reported in this paper is partly supported by
National Natural Science Foundation of China (NSFC) under No. 61300042 and No.
61321064, Shanghai Knowledge Service Platform Project No. ZF1213.
Search WWH ::




Custom Search