The role of the data management model

The role
of data management model Data management is inseparable from the data management model. It can be said that for people engaged in data management, whether he is a party or a service party, the data model has always been the focus and difficulty of data management. It can also be said that there are as many data management models as there are data management systems. Although some foreign standardization organizations provide some petroleum data management models, there are very few data management systems that are truly based on these models. Many data models that claim to follow certain standards are half-concealed and mixed with many of their own. Recognition and "special circumstances". The data management models of various domestic oil companies and oil fields are more colorful, and each has its own characteristics.
Of course, every data management model is the crystallization of the wisdom of the project implementers. It has its rationality and scientificity. When considering it from a purely technical point of view, it is often impossible to judge the pros and cons of a data management model. Many data management models The problem is gradually exposed in the data application stage, but by this time, many things are a bit late.
In terms of the range of data included, the model can be large or small. The biggest is the integrated management project of oil exploration and development data implemented by various oil companies in recent years. In these projects, it is hoped to establish a model covering all the data of oil exploration and development. , We can put all the data in our production, research, and management activities in this huge model, which can be managed and maintained by the headquarters. Regardless of the quality of a specific data management model, from the basic concept of the data management model, a reasonable data management model still needs to follow some basic principles and meet some basic requirements.
If you think about it simply, the data management model is actually a structural model for storing or managing all the information of a type of business data. No matter how much data a model contains, a simple criterion to measure the quality of a model is to see whether the model contains all the information that the data should have. In the " Essence of Oil Data"In this article, we mentioned that a complete description of data should include seven elements of data (object, attribute dimension, data representation dimension, data generation dimension information, data comment dimension, data usage dimension information, Data management dimension information), plus nine aspects such as data relationship and data source information. A good model should cover these elements as much as possible. This is the basic requirement. Specifically, there are the following requirements :
1) Store all the data information in.
All the data information is actually the seven description dimensions of the data plus the data relationship and data source information. A good data management model should be able to put all this information in. Here Among the several aspects of information, if divided from a macro perspective, it can basically be divided into two categories. One is the basic information of the data itself, mainly the content of the object and attribute dimension information, and the other is other information outside the data body. Collectively referred to as metadata information. At present, most data management models only put the first part of the information in, and only a small amount of metadata information is reflected in the data management model.
Due to the high degree of uncertainty in the understanding of petroleum exploration and development, in the data In the process of use, metadata information is more important than the information of the data itself. It can be said that oil exploration and development data lacking metadata information is basically useless data. So from the perspective of application, include metadata information in the data management model Very necessary.
2) Preserving the professional logical relationship between
data The main body of petroleum exploration and development data is to describe all aspects of the attribute information of the underground geological body, and the different data are given a strong professional logical relationship at the beginning of the design. This kind of logical relationship plays a very important role in the use of data. In particular, underground geological bodies are described by indirect information, and it is necessary to combine and explain the data of various dimensions, which proposes a more important role in the management of the relationship between data. High requirements.
The professional logical relationship between data is difficult to directly manage, and cannot be directly identified, and it is impossible to automatically or manually establish the logical relationship between the two data. Because the oil exploration and development data is actively generated data, The relationship has been established at the beginning of designing the data. In the data management model, the relationship between data can be completely managed by recording and managing the relationship between data targets, attribute dimensions, etc. in accordance with professional design ideas.
3) It can cover that the entire business
data is generated in the business, and it also describes and records the business process and business results. A good data management model should be able to cover the entire business process and record all the data generated during the business process. In order to achieve this goal, various oil companies have implemented many large-scale data management projects in the past ten years, trying to store all the data of exploration and development in certain large-scale data management models. Store and manage in the model. The main reason is that from a business perspective, data has multiple dimensions, such as dynamic data, static data; field collection data, research results data; headquarters data, branch company data, and basic unit data; data from different lines of business, such as exploration data, Development data, reservoir data, production data, etc. A simple model is difficult to include data of different dimensions.
4) The ability to simply expand the
petroleum exploration and development technology according to a model is constantly developing. A new technology application will generate a batch of new data. People definitely hope that the data generated later can be immediately integrated into the past data management system. It is required that the designed data management model is easy to expand, and can be expanded in accordance with a unified model to meet various unknowable needs in the future.
5)
The ultimate goal of data application personnel to understand and apply data management is application, and it is professional application. For professionals, their application data is based on professional habits to think and propose requirements, such as I want XX For the well location data, he actually wants to see the well location measurement data, as well as the well location map. At the same time, he may also want to see the original well location design report. In addition, in the process of research, it is necessary to obtain "relevant results and basic information about the reservoir research in this area in the past five years". When they put forward these requirements, it was completely professional logic and thinking. They did not know or care about these data. In what structure to store it, and where to store it. The data management model must be able to identify these business meanings and be able to find the corresponding data according to the business meanings.
6) Design ideas of data management model
From the perspective of the full life cycle of data, including the three functions of data storage, management and application, the data management model we designed should theoretically meet these three requirements. At present, the conventional model hopes to use a data management model to unify the storage, management and application of data. This is theoretically reasonable. However, because the storage, management and application of data involve completely different technical requirements, a unified It is almost impossible for the model to meet the needs of the three aspects. It is also proved that the various unified data management systems that various oil companies have spent huge sums of money to establish are difficult to play a role in practical applications.
Let's refer to the development ideas of the commercial Internet. In the past, traditional shopping malls stored goods by themselves, managed goods by themselves, and provided direct sales to customers. Like our current data management thinking, storage, management and service are integrated. Taobao and JD have adopted new management models, which overturned the traditional business model. In essence, it separates the storage, management and service of the goods, and hands the storage of the goods to the merchants (part of it themselves), and the management adopts a virtual management interface. Unified cataloging and management of all commodities, and adding a large number of commodity information and user usage information that users care about, and the service is handed over to an independent express delivery system. Through the separation mode, not only can all commodities be placed on the system (in theory), but also a unified and consistent management method and a more flexible and unified service method can be provided.
In the past two years, we borrowed the idea of ​​commercial Internet and designed the OiO data center . Its essence is to separate the storage, management and service of petroleum data. The focus is on the design of a unified data management model and implementation technology to achieve unified management of all petroleum data. And service (pictured below).