Cloud-based Scientific Workflow Scheduling Algorithm

Many complex scientific applications are modeled in the form of workflows to carry out large-scale experiments. Because of complexity of these applications, they need a high performance computing infrastructure. Cloud is a new paradigm that delivers computing services and enables scientists to run their application. We propose an algorithm to schedule scientific workflows on cloud resources, based on scientist’s preference QoS. After ranking the tasks, the algorithm use a FF function to select the fittest resources to gain QoS. We experiment with a workflow application by varying data size and computation cost. We compare the cost saving when using proposed algorithm and HEFT algorithm. In the experiment we consider user more cautious for cost. For this case, our preliminary results show about 12 percent cost saving comparing to HEFT.

کلیدواژه ها:

component ، scientific application ، workflow scheduling ، cloud computing ، quality of service