DMGrid: A Data Mining System Based on Grid Computing
ADMA '08 Proceedings of the 4th international conference on Advanced Data Mining and Applications
| Hi-index | 0.00 |
Recently, grid computing has become a popular technology for large-scale scientific or business computing. Grid workflow system can enhance the flexibility and efficiency of the grid computing. By the virtue of the dynamic and heterogeneous nature of grid environment, grid workflow system requires a resource evaluation and utilization enhancing mechanism to satisfy all kinds of grid resource requirements from workflow instances and ease to build billing system. Motivated by this, we present a novel Extended Temporal Logic based Workflow Specification (ETLWS) model to meet with the requirements above. The ETLWS extends temporal logic based workflow specification model by injecting resource requirement descriptor. Then, we propose a dynamic workflow scheduling scheme to improve the stability and efficiency of grid workflow system. In particular, we implement a system prototype and present a case study. The experimental results demonstrate the effectiveness of the ETLWS model and its corresponding scheduling scheme.