Hybrid databases coming quietly (1)

The database and application software branch of the China Software Industry Association was established at Peking University in March 16. This will attract the attention of national securities, insurance and other industries and database vendors such as IBM and NCR.

After more than 40 years of development, the database has gone through the first generation of hierarchical databases, the second generation of relational databases, to the debut of XML, and then to the market of object-oriented technology. However, databases, as the hardest core and strongest bastion in the software field, are facing the impact from all directions, such as the unblocking of Open Source Trends and the structural structure of databases on the Internet ...... How to deal with these challenges is also what database and application software branches and many database vendors need to do.

For the vast majority of IT personnel, the database is equivalent to a relational database, data is closely linked with the table, the E-R model has almost become the only way to describe the world, SQL language is the only means of database information access processing. Relational databases have become a religious belief. Tracing the database's short history can help us sort out the future of the database.

"Database" was first used as a phrase in June 1963, and System Development Corporation sponsored a project entitled "Developing a Computer Management Center Database" "Development and Management of a Computer-centered Data Base "). Databases, as an independent word, appeared in Europe in the early 1970s s and were not used by major American newspapers until the end of the 20th century (Data Bank Databan ).

The data model is the core and basis of the database system. Therefore, the division of the database technology development stage should be based on the development and evolution of data models. In general, the database technology has gone through three stages of development from the beginning: the first generation is the mesh and hierarchical database system, and the second generation is the relational database system, the third generation is a database system featuring an Object-oriented Data Model.

For the first time, hierarchical databases separated Data Management from application logic. In 1970s, Dr. E.F Codd proposed the relational theory, and the SQL language appeared at the same time. The second generation of database systems supports relational data models. Relational Models not only have the advantages of simplicity and clarity, but also use relational algebra as a Language Model and relational data theory as the theoretical basis. Therefore, relational databases are characterized by good forms, strong data independence, and non-procedural database languages. These features are significant signs of the development of database technology to the second generation.

Although the relational data model describes the structure of real-world data and some important correlations, it is still not enough to grasp and express the rich and important semantics of Data Objects. Therefore, it is a syntax model.

In the middle of 1990s, XML emerged as a data storage format. It organizes data in a hierarchical manner and can clearly describe complex objects and be easy to expand.

XMLExtensible Markup Language) technology plays an increasingly important role in the current IT environment. In fact, IT has become the basis of data exchange standards and SOA architecture, microsoft Office 2007 will also store documents in XML format. I recently tried this product's Beta2 version, it is found that documents stored in XML format save 2003 or more storage space than Office 20% ). According to a report released earlier by IDC, XML technology will be widely used.

At the same time, XML also blurs the strict line between traditional structured and unstructured data, impacting the development of relational databases over the past 20 years.

Object-Oriented Technology encourages designers to use objects to think about data. There is a classic analogy: You want to park your car in an object-oriented garage. Drive the car into the garage, get off, close the door, and return to your room. When you want to go out, you just need to enter the garage, get into the car, start, and drive away.

You want to park your car in the garage of a relational database. Drive the car into the garage, get off, unload the doors, put them on the ground; unload all the wheels, put them on the ground; unload the bumper and other things. Then return to your room. When you want to go out, enter the garage, first install the door, then install the bumper, and then the wheel and so on, all finished, drilled into the car, ignited, and then drove away.

In today's application field, object-oriented design ideas and technologies have been widely used in software development, but the object-oriented database technology has not been accepted by the market.

When the emergence of object databases, relational database providers have accumulated a huge momentum and occupied a large market share. In the standard SQL interface, object-oriented programs that access relational databases are easy to write. However, most of the earlier object data does not provide SQL interfaces, and it is not suitable for any query applications.

The emergence of XML and the formulation of relevant international standards, such as Xquery and SQL/XML, provide a new impetus for the introduction of object-oriented technology to databases. Although relational databases are driving the company's business development, they cannot compare with the semi-structured content represented in XML. It has invested a lot over the past 20 years, enterprises have invested trillions of dollars in relational data and relational database management system (RDBMS.

Future debate

At the expert seminar organized by Software Development, Don Chamberlin is confident that XML will be at the center of future databases: "XML will become the main format of document exchange ."

Cattel has a different view: "I don't think a pure XML database will have a bright future. Do people store and query data in XML format completely? Performance is definitely a big problem! I stood on the other side 10 years ago and agreed to solve the problem through object-oriented technology. In fact, mainstream vendors have already supported objects well in relational databases. Interoperability and compatibility are of course important. XML is a common data exchange model. Due to its scalability, it may be very suitable for specialized markets for various special purposes, but it will never replace relational databases ."

Don Chamberlin also admitted that XML is flexible and self-describing. Many programs need to query in the format of exchanged data. All other data will be mapped to XML and connected to the outside world through interfaces. However, relational databases and SQL do not exit the stage of history because they have a natural advantage in terms of data independence.

It is not advisable to simply replace RDB with XML storage. The world can be relational or hierarchical. The two have their own advantages and disadvantages. Where is the future of database technology?

Database experts said, "There are two solutions: one is to improve XML support for existing relational databases, the other is to simply use XML as the data standard to develop a pure XML database." However, developing a new database fully designed based on the characteristics of the XML data model is not only difficult but will inevitably lose the advantages of commercial relational databases in data storage.