Research On Grid-based Postal Rates Calculation Task

The goal of Grid is to connect all kinds of distributed and heterogeneous computation resources together by the high speed network for solving large-scale application problems in coordination, realizing distributed sharing of wide area information resources, finally conforming the entire Internet into a super virtual computer and realizing comprehensive sharing of computation resources, memory resources, data resources, information resources, knowledge resources and expert resources. It will be the third information technology revolution after Internet and WEB. It is rare to use grid computing technology in postal system in the present. Building a post grid platform can easily solve the problem of distributed, heterogeneous, large-scale resource sharing in postal system, as well as collaborative computing.This dissertation analyzes the prospect of massive data management in current postal system using grid technology, and make the example on the demand of postal price calculation task, put the grid into the field of postal rates calculation, I briefly state my researches as follows:First, put forward a distributed parallel computing model. This model makes full use of the advantage of postal grid to overcome the problem of overload in central computing node and large-scale data remotely transfer, which caused by the traditional method of Postal Price Calculation, distribute the calculation tasks to each bureau node and execute them parallel, minimize the calculation task finished time.Second, according to the characteristics of PPC (Postal Price Calculation) task, this dissertation puts forward a task decomposition method, which is decomposed by the combination of region and time period; furthermore, analyzes the PPC decomposition system architecture and brings in an idea of two level task-pool management mechanisms.Third, considering the aim of calculation task scheduling, the basic service of postal grid and the characteristics of decomposed calculation tasks, this dissertation puts forward a dynamic scheduling Fast-Balance algorithm. This algorithm can dynamically adjust each node's load, according to the state information returned from grid and running job monitor service, at last make all the bureau nodes load balance.Finally, we will use the research method and computing model in postal grid system project. Through decomposing the calculation task, scheduling and aggregating the results, we can solve the large-scale calculation problem by executing the distributed tasks parallel, and then enhance the efficiency of task execution, and minimize the task finished time.