At present, enterprises have been developing more and more quickly in China. In order to strengthen the systems of enterprise management, manufacturing execution, equipment and so on in each link of management, many enterprises have started to use data integration to construct an Informationization management system. Based on the needs of the development of enterprises, this paper analyzes the pattern formation of data integration and its advantages and explores its role in enterprise informationization construction. Introduction In the times of informationization, informationization development is a necessary stage for every enterprise to go through in metamorphosis. In the process of enterprises’ development, it is easy to form “information island” due to failure of communication or trade restrictions, which troubles either decision-making or management for them. Latitudinally,enterprises need to receive the latest industry news for development; longitudinally, they need to summarize experience of historical development; on the whole, they need to integrate information from each subsumed department. In all, informationization is necessary, so is data integration for enterprises’ development in terms of the three above-mentioned aspects. Data Integration The so called data integration is divided into three layers in the light of its structure, i.e., data-capture layer, data management layer and data service layer, as is shown in Fig. 1 below. The data-capture layer is responsible for capturing and updating available data and then registering them as data source. The data abstraction model, namely, data-capture layer, first monitors and analyzes data from data sources in storage. As an enterprise embodies a big database, the basis of model construction relies on data collection. Therefore, it is necessary for the data extraction model to detect changes in data for adding in data at all times. The data management layer is responsible for data integration, i.e., management and maintenance on the basis of data. As the data is in a scattered state after being collected, only when they are integrated and classified will they be available for backup use. The data service layer provides useful information for decision makers pointedly and selectively. However, decision making calls for precise and accurate data. Hence, the provided information is supposed to having been analyzed in one way or another to the necessity of decision making. All the three layers are constituents of EAI(Enterprise Application Integration). They seem to perform different functions, but constitute an inter-related and indivisible whole. They are all described in XML because the language is both easy to use and has good performance. Data International Conference on Computational Science and Engineering (ICCSE 2015) © 2015. The authors Published by Atlantis Press 223 extraction, transfer and loading are all components of the establishment and maintenance of database, for which several key techniques are adopted including data extraction model design. Figure 1 Layer Diagram of Data Integration System The data load design, for example, is one of the core technologies employed for data integration. At present, inverse transformation of data is carried out with the help of XML. That is, XML data is transformed into target data; it is also a process of loading data into database. In the process of data loading, target increment table is a key point as well as the source of data to be loaded.The names for the target increment table may includean extracted table name with the suffix “-dest”or flag, updatetime, etc. addingall of the fields of the corresponding target. In general, constructing data loading first requires connecting the database to the target database and then the data in the target increment table load to the target database.That is the process of data loading design, wherein the loading method employed for the target increment table as the data source can import data of similar structure without switching in the context of the database. This greatly improves the performance of data loading. Necessity of Data Integration Nowadays, informationization is becoming more and more important. In addition to capital and technology, information has become an important factor for the development enterprises. Therefore, database technology has become the hot spot of research and received as much attention as in the research both at home and abroad. But most of the concern has been focused on technology and integration framework, which has changed related methods of data integration and is of great help for enterprises to improve flexibility and collaboration capabilities. It is widely acknowledged that database has good real-time performance. But it has certain requirements for network and devices and its cost is high. If the transformation and transmission is processed through data replication, even though the data could be delayed, the access will be faster and cost will be lower. Therefore,it can be concluded that although data integration has faults, the methods for its implementation are simple. Besides,it is flexible in s data processing andhas good maintainability; thus feasible. In the meantime, it is also an "informationization revolution" necessary for enterprises to carry out to stand in the tide of information in the information age.