Research Of Workflow And Mass Data Storage In Model Management System

With the high-speed development of information technology and the increasingpopularity of the Internet, corporate management on the models has also undergonetremendous changes. The management during the design process of the product willgenerate large amount of models and the process control is complex so that needing apowerful model management system to achieve the storage and management of themodel. Traditional single version model management system only provides limited datastorage capacity and processing power, the scalability and fault tolerance are also poorlead that it is difficult to meet the complex needs of large model management for largeconglomerates.In response to above problems, the model management system takes advantage ofB/S mode and three-tier architecture to provide users with a friendly user interface andhigh cohesion and low coupling system architecture. Because that the process control oflarge systems involves large number of documents, complex process and complicatedaccess control relations, designing workflow engine based on Petri nets to execute theprocess control function. The edge is on behalf of an operation in the Petri net and thenodes represent the beginning or the end of the state of certain operations. Then proposea data storage solution for the separation of the metadata and data in the model to meetwith the request of data storage and management for mass model. Metadata is structureddata and is stored in the relational database; data is unstructured and requires a lot ofstorage space, so use open source distributed file system-HDFS(Hadoop Distributed FileSystem) to bank, the key/value pairs are stored in the file-based data structure-MapFile.In addition, through the use of the relationship between table and table in the relationaldatabase tables, tree file system virtualization will be generated based on the linearstructure to supply users with an efficient file functions on the management model.Experiments show that the workflow engine based on Petri nets can make theprocess control functions in the model well, the proposed solutions for the mass datastorage and management meets the system’s request for mass data storage and management which reaches the validity for data, scalability, fault tolerance and alsoprovides an effective management style for users.