Research For Software Architecture Of Big Data Management Platform For Warning And Decision Support

Big Data application development is in full swing for these years. Sense ofabstraction-processing-prediction-decision model, which procedure containsacquiring data flow from business, managing and processing for the generatedprediction model, predicting in order to achieve early warning decisions, totallyincluding forecasting model learning and online applications. Analysis and designlarge data management software architecture and implementation is to support largedata application systems. This paper made the following tasks:1. Bring up big data management platform domain model for warning decision.For large data management, identify common needs and changes in the field of borderareas. Establish the domain model, including role model, feature model, informationmodel, business model and domain ontology.2. Bing up big data management platform software architecture for warningdecision. Proposed for large data management needs, design large data managementplatform solutions, determine the core software architecture model from five aspects,as logical view, process view, development view, physical view and the scene. Designcomponents, connectors, configuration interaction relationships and constraints, anddescript them using the Z formal language.3. Develop a big data management platform prototype system using thebackground of power plant production safety monitor. Based on the achievements ofarchitecture design, go through implementation process, implementation architectureand network topology of the prototype system, and describe multi-sourceheterogeneous data storage structures.In this paper, we made data management domain modeling and architectureguiding from domain engineering methodology. It is an important innovation in thefield for system design, research and development, and would play a catalytic role forsubsequent application development.