TIMOM: A Novel Time Influence Multi-objective Optimization Cloud Data Storage Model for Business Process Management

Abstract

In recent years, lots of the BPM (business process management) data storage strategies are proposed to utilize the computation and storage resources of the cloud. However, most of them are mainly focusing on the cost-effective methods and little effort is paid on the time influence of different datasets in the BPM. Storing the datasets with high time influence in the cloud is conducive to reduce the total cost (storage cost and computation cost) which is also important to the better performance of the BPM. Aiming at this problem, this paper proposes TIMOM, a novel time influence multi-objective optimization cloud data storage model for BPM. In the TIMOM model, the time influence model of the BPM is firstly constructed to calculate the time influence value for each dataset. Based on the time influence value, a new strategy, SHTD, for storing datasets in the cloud is designed. Then, by taking response time and total cost as two objectives, the multi-objective optimization method is designed to optimize the performance of the SHTD strategy. This method conducts the non-dominant sorting and calculates crowding-distance for all datasets. By doing this, important datasets are selected and stored in the cloud for repeatedly usage. Experimental results have demonstrated that the proposed TIMOM model can effectively improve the performance of the cloud BPM systems.