Home » Blog » What is the future of database management systems?

What is the future of database management systems?

Now that the database management system changes everyday, we are very honored to invite Professor WANG Hongzhi to conduct an analysis of the future trend of the database management system development.

This article is only a personal opinion, please bear with meif there is any bias

🤠🤠🤠

What is the future of database management systems?

 

In today's era where data is king, the database management system has become one of the most important infrastructures. In fact, it has already become a fact that informatization rapidly developed more than 20 years ago. At that time, relational databases took the lead under the goal to express real-world entities and relationships in the information world. Oracle, which focused on relational database management systems, and Microsoft, which included the entire desktop product line including operating systems, programming languages, database management systems, and games, were evenly matched. Although many new database management systems such as XML database and object database have been proposed, most of these new data management systems are limited to research prototype systems, and hard to substitute relational database management systems in the market.

With the advent of the Web 2.0, mobile Internet era, and especially the big data era, things have gradually changed. Due to the diversification of application loads, data types, and hardware and software platforms, the requirements for databases have changed from “one size fits all” to “no size fits all”. More intriguingly, both the “No-SQL” movement and the concept of big data, which have subversive effects on relational data, appeared in 2009.
Today, with the rapid growth of data scale, modality, and data-centric applications, database management systems diversifies. Such diversification is presented in the following aspects, some of which have been recognized by the market, while others are still under exploration.

Hardware empowerment

Storage and computing are the two most basic requirements for the computing system put forward by the database management system . Therefore, new storage and computing hardware will inevitably become the tipping point of the new database management system. Currently, new storage hardware such as Flash and NVM and new computing Hardware such as GPU and FPGA have brought new opportunities and challenges to the database management system. With the continuous development of materials, electronics and other fields, it is believed that various high-performance hardware will continue to appear, which helps create higher-performance database management systems.

AI empowerment

In current database application, there is still a lot of human participation, including parameter setting, index and storage structure selection, backup and recovery timing selection, and even database management system selection, which makes DBA very important. However, as data scale increasing and model changes accelerating, it becomes more difficult for people to understand the overall picture and changes of data, which makes the work harder. With the development of artificial intelligence, database professionals are inspired to use artificial intelligence technology to replace part of labor, realize automatic parameter tuning, automatic index recommendation, automatic storage structure design, and then realize the autonomy of the database.

More general

Due to reasons such as the independence between each department, both the data and data processing in the same organization have become complicated. Even small applications often involve data from multiple sources and the entire processing schedule. For these scenarios, traditional database management systems and data warehouses have evolved into systems such as data lakes, incorporating data integration and quality management. , storage and other functions, become a more general data processing system.

More dedicated

The data management of different modalities or even the same modal data in different applications does not require storage structure, indexing or query processing. It seems that using the same system to process different applications is somewhat inappropriate, and it is difficult to achieve extreme high performance. Therefore, some large manufacturers today tend to develop one or more sets of dedicated database management systems for specific applications, and continuously optimize performance based on application-oriented features. Currently, a series of dedicated database management systems for single-modal data, such as time series databases, graph databases, document databases, and spatiotemporal databases, etc. Databases are booming, and even time-series databases have proposed different database management systems optimized for Internet of Things, financial and other applications

Bigger

In applications such as the Internet, the scale of data is extremely large and even geographically distributed in multiple centers. It is nearly impossible to query and analyze these data intensively. Highly scalable database management systems are urgently needed.

Smaller

As opposed to the need for large databases to manage larger volumes of data, there is also a need for embedded small databases that can be linked into applications. This type of database can be embedded in the process and does not require a separate database engine. The system is customizable and small in size, which can meet the needs of embedded systems.

The burgeoning field of database management systems can hardly be covered in one article. It is precisely the basic role and various new technical requirements of database management systems in the information era that attract us database practitioners to continue keep working.

About the Author:

Prof. WANG Hongzhi is the professor and doctoral supervisor of the Computer Science Department of Harbin Institute of Technology, vice president of the Elite College, director of the Massive Data Computing Research Center, and head of the data science and big data technology major, and a young Longjiang scholar. His research direction is big data management and analysis. He has published more than 300 academic papers and three academic monographs, his papers have been cited more than 3,000 times, and he has authorized 30 invention patents. He won the first prize of Natural Science Award of Heilongjiang Province, the first prize of Science and Technology Progress Award of Higher Education of the Ministry of Education, Heilongjiang Youth Science and Technology Award, Baosteel Excellent Teacher Award, China Excellent Database Engineer, etc. He has presided over more than 10 projects including key projects of the National Natural Science Foundation of China. He is the chairman of Harbin Branch of China Computer Federation, secretary-General of ACM SIGMOD China, standing Committee Member of China Database Professional Committee, and one of ACM Data Science Discipline Standard Writing Group.