The Experts below are selected from a list of 3 Experts worldwide ranked by ideXlab platform
Jinjun Chen - One of the best experts on this subject based on the ideXlab platform.
-
Forecasting Scientific Cloud Workflow Activity Duration Intervals
Temporal QOS Management in Scientific Cloud Workflow Systems, 2012Co-Authors: Yun Yang, Jinjun ChenAbstract:As discussed in Chapter 2, Workflow Activity Duration is one of the basic elements in the temporal consistency model, and thus its accuracy is critical for the effectiveness of temporal verification and all the other related components such as temporal checkpoint selection and temporal violation handling. Therefore, an accurate forecasting strategy is required to predict cloud Workflow Activity Durations. However, it is not a trivial issue due to the dynamic nature of cloud computing environments. In this chapter, we present a statistical time-series-based forecasting strategy for scientific cloud Workflow Activity Duration intervals. The comparison results demonstrate that our strategy has better performance than the other existing representative strategies. This chapter is organised as follows. Section 5.1 gives a general introduction about cloud Workflow Activity Durations. Section 5.2 presents the specifically related work and problem analysis. Section 5.3 presents the novel statistical time-series-pattern-based forecasting strategy. Section 5.4 demonstrates the experimental results.
Yun Yang - One of the best experts on this subject based on the ideXlab platform.
-
Forecasting Scientific Cloud Workflow Activity Duration Intervals
Temporal QOS Management in Scientific Cloud Workflow Systems, 2012Co-Authors: Yun Yang, Jinjun ChenAbstract:As discussed in Chapter 2, Workflow Activity Duration is one of the basic elements in the temporal consistency model, and thus its accuracy is critical for the effectiveness of temporal verification and all the other related components such as temporal checkpoint selection and temporal violation handling. Therefore, an accurate forecasting strategy is required to predict cloud Workflow Activity Durations. However, it is not a trivial issue due to the dynamic nature of cloud computing environments. In this chapter, we present a statistical time-series-based forecasting strategy for scientific cloud Workflow Activity Duration intervals. The comparison results demonstrate that our strategy has better performance than the other existing representative strategies. This chapter is organised as follows. Section 5.1 gives a general introduction about cloud Workflow Activity Durations. Section 5.2 presents the specifically related work and problem analysis. Section 5.3 presents the novel statistical time-series-pattern-based forecasting strategy. Section 5.4 demonstrates the experimental results.
