Research On Metadata Classification Method Based On DOA

The information technology which is symbol of the 21 world brings a hitherto unknown change to human life with its development and popularization. The development of network technology results in data distributed storage in different devices flew in every moment. Existing software architecture can not adapt to the era of big data, because of its own limitations. It is difficult to solve the problem of information sharing, system expansion, efficient management and data mining and so on. In order to solve the problem of facing the era of big data, DOA architecture with putting forward new ideas emerge as the times. Architecture Oriented Data Center(DOA) registered by the data(DRC), data access center(DAC) and abnormal data center(DEC), a unified definition of management and provide data services; through the application data unit(DAUs) for the management and service of various applications, build a data ecosystem data platform with the application of fragmentation.The data registration center that is a logical resource pool storage a large number metadata which has been registered. With more and more data registration, the total amount of metadata is increasing. The problem of metadata classification, storage and retrieval are more and more prominent. The classification of the metadata plays an important role on the storage and retrieval of metadata, so the metadata classification has important significance. The metadata classification has the advantage of saving storage space, improve storage and retrieval efficiency shorten the search time. At the same time, metadata classification has a certain commercial value to the enterprise. We can forecast the industry trend, user behavior analysis, provide decision-making basis for enterprise metadata classification. Due to the widespread use of metadata, the metadata is more complex, more and more difficult to divide and more species. Therefore to explore a scientific classification method is currently urgent problems.This thesis based on DOA and metadata,combined with the characteristics of metadata study the metadata classification method. This thesis designs the process of metadata classification though multiple perspectives of the classification of metadata. According to the DOA data registration center of their own needs and comparative analysis of the current popular metadata standard, we formulate a common metadata standard; metadata field is proposed according the theme, time, location and frequency of the four retrieval rules; as soon as the metadata is classified, it will be storaged and retrieved. Therefore this thesis describes the metadata storage and retrieval problems after the metadata was classified.The main work is as follows in the thesis:(1)The metadata theory research.DRC is core of DOA, metadata is the core of DRC.(2)Establish a metadata specification. Analyses the current metadata standard and combines with the characteristics of data registration center, develops a universal metadata specification of DOA.(3)The extraction of classification rules. Through the study of existing metadata classification rules, and combined with the metadata specification is formulated in this paper, based on metadata retrieval, frequency, time, regional theme classification rules.The research results and innovation of this paper:(1)Proposes a general metadata specification. The thesis customize a general metadata specification based on the research of the existing metadata standards and combined with the data registration center demand for data management.(2)Proposes a classification rule based on metadata fields. According to the characteristics of metadata and DOA, around for storage and retrieval of the idea, this paper proposes a classification method according to the metadata field, which is in accordance with the metadata in the field of theme, time, region, frequency classification retrieval.