Xiao Liu PhD Thesis.pdf - Faculty of Information and Communication ...
Xiao Liu PhD Thesis.pdf - Faculty of Information and Communication ...
Xiao Liu PhD Thesis.pdf - Faculty of Information and Communication ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
successful completion. Specifically, in our temporal framework, the foundation is<br />
the probability based temporal consistency model, <strong>and</strong> its capability is realised by<br />
three components which supports workflow instances throughout their lifecycles,<br />
including temporal constraints setting, temporal consistency monitoring <strong>and</strong><br />
temporal violation h<strong>and</strong>ling.<br />
Figure 3.1 The Probabilistic Framework for Temporal QoS<br />
As depicted in Figure 3.1, our probabilistic framework consists <strong>of</strong> three<br />
components which can provide lifecycle support for high temporal QoS in scientific<br />
cloud workflow systems. The three inner cycles st<strong>and</strong> for the three important factors<br />
involved in the scientific cloud workflows: real world applications, scientific<br />
workflows, <strong>and</strong> cloud services. All the basic requirements for scientific cloud<br />
workflows come from real world applications. Real world applications need to be<br />
abstracted by service users with the support <strong>of</strong> workflow modelling tools <strong>and</strong> then<br />
create the workflow specifications. With the workflow specifications (usually in the<br />
form <strong>of</strong> visualised workflow templates) submitted by the service users, scientific<br />
workflows are executed by the cloud workflow systems with the underlying cloud<br />
computing infrastructures. Cloud workflow systems, as a type <strong>of</strong> platform services<br />
themselves, can utilise many other cloud services which provide s<strong>of</strong>tware services or<br />
computing services. Evidently, the execution <strong>of</strong> scientific cloud workflows is<br />
23