Recently, Yang Chuanhui, CTO of Beijing Aoxing Beisi Technology Co., LTD., accepted an interview with CSDN – [Here comes the Big man] column, discussing the product itself, the industry needs, and the future of the database.

As a leading technical expert in the database industry, Yang chuanhui made an in-depth analysis of the most concerned hot topics in the industry, gave a comprehensive answer to the construction planning and layout of OceanBase open source community, and made predictions and prospects for the future development trend of the database.

The big coffee to [is] CSDN series of high-end dialogue, with “dialogue technology higher-ups, programmers story” features, covering the frontier technology, open source, technology, entrepreneurship, and other fields, each invited industry representative and influential higher-ups, depth to the value of subject, the application of the technology development and innovation of multi-dimensional insights.

Click the link to watch video interview: www.ixigua.com/69906704064…

The following is a transcript of the conversation between Yang Chuan-hui and CSDN:

What are the types of industry databases? What are the characteristics?

Yang Chuanhui: Databases are mainly divided into two categories: relational database and non-relational database. Non-relational database includes key database, document database, graph database, geographic information database, etc. Relational database can be divided into OLTP transactional database and OLAP analytical database. The main advantages of OLTP transactional databases are the ability to handle high concurrency and ACID transaction processing capabilities; The advantage of OLAP analytical databases is that they can handle complex queries and have the ability to execute large queries in parallel.

Today, some mainstream OLTP databases, including Oracle and SQL Server, have strong OLAP analysis capabilities. There is a relatively new term called HTAP, which refers to the database that can process both OLAP and OLTP.

How have data consumption habits and data architecture changed in the last decade?

Yang Chuanhui: It is often mentioned in my share that application is the first driving force of database. In recent ten years, two major changes have taken place. The first change is the rise of Internet and mobile Internet, which makes some application scenarios of database become open scenarios from closed scenarios. For example, in ATM withdrawals, we have as many atMs as the number of users; In today’s Internet era, people shop and pay online and offline. Especially on The Singles’ Day, there will be hundreds of millions of people in China doing highly concurrent operations at the same time. The second change is that China’s database users will pursue independent innovation, many enterprises need to be able to control the database.

From the perspective of data architecture, there have been two changes. The first change is that application requirements have driven the transformation of database architecture from centralized to cloud-native and distributed. Through the distributed architecture, there are more common servers built into a distributed cluster, so as to meet today’s Internet, mobile Internet trend under the high concurrency and scalability needs; The second change is that the data processing type has developed from simple structured data to richer multi-mode data and unstructured data processing, and document database, key value database, graph database and other richer database types have emerged.

What are the “pain points” for Chinese database users? What is the solution?

Yang Chuanhui: For China’s data, there are two pain points. First of all, from the perspective of products and technology, China’s database is facing the transition to cloud environment and distributed architecture. The main pain point is that the classic centralized database has no way to handle the larger amount of data and no scalable capability. And because the classic database relies on high-end hardware, servers, storage, so the cost will be relatively high;

The second solution is the independent innovation capability of China’s database. There are two main solutions: the first is the secondary development based on the open source database, and the second is the complete independent research and development from scratch. Most companies will adopt the first solution, but only OceanBase chooses the second solution: the 0-1 complete independent research and development.

Then based on the open source solution, the advantage is that the initial investment cost is low, because there is an open source database as the basis, can quickly start; However, the self-developed scheme will be more controllable and have greater potential in the later stage. OceanBase chose the harder option because we wanted to do enterprise-level databases.

Advantage is the ability to open source database query processing is simpler, but there is no way like enterprise database has rich complex query ability, nor has the ability of distributed, to complex queries and distributed, we think the kernel must break open source database, full control kernel, to put this matter perfectly, So OceanBase went from 0-1 to fully independent.

Looking back today, we find that it was a good choice, because since 2017, when Ant Group realized all of its core business from Oracle to OceanBase, we found that our capabilities have surpassed the open source database, and as we have more and more control over the database, It is also constantly opening up the gap with open source databases.

HTAP is on the rise. Will it replace relational databases, such as Oracle, or traditional data warehouses? What is the future of databases?

Yang Chuanhui: First of all, HTAP is not a new concept. When databases were first created, there was no distinction between OLTP and OLAP. Today, when we look at very successful commercial databases, such as Oracle or SQL Server, they can handle both OLTP and OLAP. They can also be considered as HTAP databases. Of course, with the introduction of cloud and distributed technologies, we have expanded the boundaries of HTAP through distributed architecture. Classic databases can only be used in centralized scenarios and can only handle relatively small amounts of data, without scalability. The new generation of distributed database is capable of scalability, higher concurrency, and more elastic capabilities. It certainly extends the boundary of previous classical database processing, and can handle massive data.

The future is the era of the cloud, I think that the next generation of database, it must be adopt the distributed architecture of a new generation of HTAP database, on the one hand, such a database, it is the underlying native distributed architecture, able to fully enjoy the native dividends distributed technology, such as extensible ability, have the ability to high concurrency, high cost performance;

On the other hand, the next generation of the database must stand on the shoulders of the classical database giants, fully draw on the classical database to do some of the better technical solutions and design concepts, such as draw on the classical database SQL compatibility, transaction processing ability and other capabilities; Learn how to do refinement, how to do storage and computing separation of technical solutions. Only by standing on the shoulders of classic database giants and integrating the distributed technology into the next generation, can we make the next generation distributed HTAP database that conforms to the needs of users.

What are the planning and layout of the OceanBase open source community?

Yang: I think developers should choose the most promising development projects, and OceanBase is destined to be a top open source project.

First of all, OceanBase’s core technology has been widely proved by the industry. OceanBase’s technology in the distributed field is the world’s leading, which has been proved by ant Group’s application scenarios and customers from various industries such as finance, operators and government. OceanBase is very stable, mature and reliable. When we open source such an enterprise-class database, it is destined to be successful. Moreover, OceanBase has obtained more than 2000 STARS on the first day of open source, which should be the most in the world among open source projects of database.

Secondly, OceanBase will invest a lot of energy to build the open source community, so as to make our open source community more popular. As long as our developers grow up with OceanBase, I believe that every developer will gain the growth of technology and experience. Now the open source project of OceanBase is still in its early stage. Frankly speaking, if you join OceanBase at this stage, it will be easier to find the problems of OceanBase and make contributions to OceanBase, so as to become a Contributor of OceanBase. As committer, the governance mode of OceanBase community borrows from the mode of Apache open source project, including PMC organization, contributor, committer, etc., which is similar to the general Apache open source project.

Whether an open source community can operate well depends on the sincerity of the open source project. It is obvious that OceanBase has applied the core technology of Ant Group in the business very maturely and completely opened the kernel, which is a very sincere thing. In addition, OceanBase will quickly patch all commercial modifications to a branch of open source in the future. OceanBase company itself is also a participant in the OceanBase community, and of course, it will also be the participant with the biggest contribution.

Second, the operation plan for users is divided into two parts: Head for the customer, there will be ten spring breeze of the VIP plan, this stage to join OceanBase users of open source project head, there will be a dedicated support team and service staff, for each user to provide VIP service, needs to feedback more quickly, to enable users to use OceanBase better; For other users, there will be OceanBase Forum, wechat group, Nail group and other service modes to support and answer your questions;

The third is a developer-run program, with events for developers, colleges, database enthusiasts, etc., such as monthly meet up, monthly live streaming, and in September there will also be a database specific OceanBase competition.

What stories do you and Programmer magazine have to share?

Yang Chuanhui: When I went to university in 2002, there were few technical magazines. I read every issue of New Programmer magazine carefully. I used to focus on technical analysis and technical architecture, but at this point I prefer to read more macro stuff like Qi Lu’s article in the last New Programmer magazine and the SaaS industry.

The above is all the interview content, hope to inspire you, but also welcome you to in-depth discussion and exchange.