E-government is of great importance in the construction of China's information infrastructure. For a comprehensive e-government system, its transaction processing will inevitable based on distributed work flow.Due to the complexity of events in E-government, the system tasks are extremely unbalanced in aspects like locations timing and process.Therefore one of the major challenges in E-government is to achieve process coordination and load balance in distributed systems consist of various work flows and different IT structures of different government departments.This thesis analyses typical e-government application systems, the fundamental methods adopted by load balance techniques of existing distributed workflow engines, and also analyses different working modes and load balance techniques adapted by various distributed workflow engines, compares advantages and disadvantages of different load balancing policies. Based on these analysis, the author designs a simple and feasible scheme, around an original concept, that is to avoid worse output under worse situation. And the scheduling models and policy principles based on this scheme achieves effective load balancing in a real world e-government distributed workflow engine application.The new scheduling algorithm used in this new engine is tested and proved more effective than round robin algorithm in load balancing under the situation of complex structure and multiple tasks. The new scheme is capable of significantly increase the processing capacity and speed of e-government system and reduce hardware maintenance cost. |