Skip to Main Content
This paper presents the research work in designing the architecture of translation server of Internet based machine translation system by using the infrastructure of grid. Where the translation server is able to access a LAN (or WAN) whose spare computing resources could be employed to accomplish the massive translation works generated by overwhelming user requests. According to the characteristics of different analytical tasks in a translation system, the whole process can be factorized into a series of tasks, where some computational tasks can be decomposed into highly independent and making the processing highly parallelizable. Two different configurations of distributed tasks over the network environment have been constructed in our current study.