The little ant says:
On April 16, 2018, the US Department of Commerce banned Chinese companies from selling parts to Chinese telecom equipment maker ZTE corp for seven years.
In effect, all U.S. companies are prohibited from doing any kind of transaction with ZTE. Because zte’s communications products require core components imported from the United States, some industry critics say this almost means the company is doomed.
Following the announcement, ZTE’s shares were suspended on both the Shenzhen A-share market and the Hong Kong stock Exchange, an unprecedented level of suspension.
After the continuous outbreak of “ZTE incident”, the voice of basic software made in China gradually rose. Among these voices, the author found a voice different from the past, and many people praised and supported it: “The localization of basic software can only be successful if Internet companies do it.”
Although, this voice may be a little idealized, but it also shows that everyone is very recognized for the technology and self-research ability of Internet companies. Ant Financial as a well-known Internet company, its technical level has been talked about by the public.
Ahead of the upcoming DTCC 2018 conference, we caught up with Ke Feng, senior director at Ant Financial and chief architect at OceanBase, to get his thoughts on homegrown technology.
【 Introduction: 】
Before joining Ant Financial in 2014, he worked as CTO in Shenzhou General, a database manufacturer. He has a doctor’s degree in computer application from Zhejiang University and 15 years of experience in database development and industrialization. Currently, he is working as an architect in the Basic Data Department (OceanBase team), participating in the design and development of OB 1.0. His main research areas are distributed relational databases, data storage, performance diagnostics and optimization.
1. There is no good OR bad technology. Both open source and self-developed technology are business drivers
Before we based on interviews with domestic software makers, hear a lot about open source, has said, “open source is not controllable, innovation rely too much on open source easy subjugation”, also somebody said “many vendors on the open source framework for innovation, and then take the market by low price strategy, caused loss of industry foundation of original technological innovation power”…
With all this, we see a lot of people who are not very friendly to doing research and development based on open source. Feng ke compared open source research and development to driving a jeep and self-developed technology to building high-speed trains. He believes that there is no absolute right or wrong choice of technology. The choice of open source or self-developed technology depends on whether it can truly solve business problems.
Oceanbase is a database completely independently developed by Ant Financial. Why did ant Financial choose self-research instead of open source development when choosing technology? “A big reason is that the distributed database is a new technical challenge and there is very little that can be reused from existing open source technologies,” Feng said. “Another important reason is technical idealism. We don’t want to be constrained by the existing open source technology framework.”
2. “I know how hard it is to develop your own technology, but the value of doing it is that we do it.”
Feng Ke is a veteran rooted in the software and Internet industry for 15 years. He has witnessed the growth of China’s software revenue from 32.5 billion yuan in 1998 to 5.5 billion yuan in 2017. Meanwhile, he knows clearly that in the core CPU, operating system and database, we have not made significant breakthroughs in the commercial sense in the past. Feng Ke knows very well how difficult it is to develop self-developed technology of basic software, but he also firmly believes that the value of self-developed technology is worth doing so.
The development of domestic self-developed software products has been a lot of twists and turns, first by pirated commercial software plunder, and then by open source ecological set. Based on open source, we can improve our technical level quickly, but at the same time, our thinking is bound to a certain extent, while self-research technology provides us with the source power of basic innovation.
Another benefit of developing your own technology is that you can achieve differentiated value. Taking Oceanbase as an example, the concept from the very beginning is to achieve differentiated value based on the core financial business scenarios accumulated by itself, so the architecture design of Oceanbase is completely different from that of Oracle and DB2. The distributed system of Oceanbase is mainly to achieve high reliability of the system through system architecture and software. Now, Oceanbase not only bears the weight of all the core business of Ant Financial’s “Double 11”, but also has landed in six commercial banks.
3. To engage in self-research, the great advantage of traditional manufacturers may be a big burden, while Internet enterprises can carry on the battle lightly
“Zte incident” is more like a big wave swept by sand, the voice of domestic basic software was completely released at the beginning of the article we wrote that many people have high hopes for the basic software development of Internet enterprises, why this situation? What are their advantages over traditional infrastructure vendors?
Before joining Ant, Mr. Feng worked as CTO at database maker Shenzhou General. He, who has been to traditional database manufacturers and is also engaged in research and development in Internet enterprises, is naturally the best answer to this question. “Established database companies have a complete ecosystem, which is a huge advantage. But a big advantage for established database vendors is also a big burden.”
In the past, domestic research database, in fact, is in a chase, imitate the development of commercial database companies, really can make differentiation of things very few, and now do database from the technical thinking has changed, from centralized database technology form to distributed technology form.
And this one change, let the ecological advantage of traditional manufacturer may become burden instead. As we all know, when traditional vendors did databases, PC server reliability was very poor, and there was no recently developed highly available, highly reliable distributed architecture. In this case, traditional manufacturers want to innovate and self-research is like turning around a big car. There are many aspects that need to be taken into account, while Internet enterprises can pack lightly.
4. To summarize
It is known to all that it is difficult to develop self-developed technologies, but Ant Financial has never been afraid of challenges in its independent research and development and innovation. It has made technological innovations in alipay guaranteed transactions, risk identification, intelligent customer service, and even more basic databases and middleware.
In the upcoming DTCC 2018 conference (May 10-12, 2018, Beijing international conference center), in addition to Feng Ke keynote, tells the story of behind the research of distributed database, other ants gold clothing technology research and development experts will bring more distributed transaction processing, large-scale figure calculation, data analysis and case sharing in the areas of application.
China Database Technology Conference (DTCC) is the largest and most popular technology exchange event in the field of database and big data in China, held annually.
Since 2010, the conference has invited more than 100 industry experts to share hot technical topics, providing the most valuable communication platform for database crowd, big data practitioners, Internet people and industry related people. The 9th China Database Technology Conference (DTCC2018) will be held on May 10-12, 2018. In this conference, Ant Financial will bring the following themes to share:
DTCC main Forum
1. DTCC and OceanBase have grown together along the way
[Lecture Time and Introduction: 2018/5/10 08:40-08:45]
Since 2015, OceanBase team has introduced database talents from various channels. Through continuous efforts, and finally realized the reconstruction of the database, and launched a truly distributed database facing the cloud era. Since 2017, OceanBase has started a practical process of empowering traditional finance. At the end of 2017, OceanBase was officially launched in The Bank of Nanjing. OceanBase has also been exported to India and the United States. At this point, OceanBase has truly realized the transformation from r&d practice to formal external commercial use. Looking back over the past 9 years, China Database Technology Conference, as a sharing platform, faithfully recorded the key R&D and application process of OceanBase and witnessed the growth process of OceanBase.
[Guest introduction: Yang Zhenkun, senior researcher of Ant Financial, founder of OceanBase independent research and development database]
In 1999, he became one of the first Yangtze River Scholars of Peking University. He was awarded the First Prize of National Science and Technology Progress (ranked fourth) and is the inventor of more than 20 invention patents. He has served successively as deputy director of The Institute of Computer Science and Technology of Peking University, chief researcher of Lenovo Research Institute, chief researcher of Microsoft Research Asia, and senior scientist of Baidu.
2. Development, practice and future of self-developed technology in Ant Financial
Time and Introduction: 2018/5/10 13:30-14:10
As a independently developed financial level distributed database, OceanBase has steadily supported Ant Financial’s Singles Day peak traffic over the years, and set a world record of 256,000 payments per second in 2017. As a new generation of relational database, OceanBase has solved a series of worldwide technical problems in scalability, high availability, high performance, low cost and other aspects, and provided a number of core capabilities for upper-layer applications, such as “continuous shrinkage and expansion”, “great promotion of flexibility”, and “multiple locations and multiple activities”. Today, we reveal the secrets behind OceanBase.
[Guest introduction: Feng Ke, Senior Director of Ant Financial, Chief Architect of OceanBase]
Before joining Ant Financial in 2014, he worked as CTO in Shenzhou General, a database manufacturer. He has a doctor’s degree in computer application from Zhejiang University and 15 years of experience in database development and industrialization. Currently, he is working as an architect in the Basic Data Department (OceanBase team), participating in the design and development of OB 1.0. His main research areas are distributed relational databases, data storage, performance diagnostics and optimization.
Second, financial core basic technology application practice special session
1. OceanBase: Financial level distributed database independently developed by Ant Financial
Time and Introduction: 2018/5/10 13:30-14:10
As a independently developed financial level distributed database, OceanBase has steadily supported Ant Financial’s Singles Day peak traffic over the years, and set a world record of 256,000 payments per second in 2017. As a new generation of relational database, OceanBase has solved a series of worldwide technical problems in scalability, high availability, high performance, low cost and other aspects, and provided a number of core capabilities for upper-layer applications, such as “continuous shrinkage and expansion”, “great promotion of flexibility”, and “multiple locations and multiple activities”. Today, we reveal the secrets behind OceanBase.
[Guest introduction: Chen Mengmeng, senior technical expert of Ant Financial]
Currently, I am in charge of SQL related development work in OceanBase team. He graduated from Tsinghua University in 2006. From 2006 to 2008, he was responsible for the development of grid computing scheduler at CERN. In May 2009, he received his master’s degree in computer science from University of Wisconsin-Madison, USA.
2. Distributed transactions: practice and evolution in core financial scenarios
Time and Introduction: 2018/5/10 14:10-14:50
Performance and data consistency is a constant challenge in financial trading scenarios. After years of double Eleven promotion and New Year red envelopes, Ant Financial has a very complete and in-depth theoretical understanding of distributed transactions in extreme performance scenarios, and has made many innovations and formed a complete set of distributed transaction solutions. This speech mainly presents the technological evolution process of Ant Financial distributed transaction solution, including in-depth thinking on transaction consistency, limit performance optimization, disaster recovery, smooth expansion and scaling, complexity of business access distributed transaction framework, scheme versatility and other aspects.
[Guest introduction: Erudite Yin, senior technical expert of Ant Financial]
Currently, I am in charge of data middleware technology in Ant Financial. Previously, I was responsible for database kernel and cluster technology direction in Baidu. He has in-depth research and abundant engineering practice in the fields of distributed transaction, database high performance/high reliability architecture, database kernel and so on.
3. Evolution of Ant data analysis platform
Time and Introduction: 2018/5/10 15:10-15:50
Ant data analysis platform, with the mission of realizing the “democratization of analysis” of Ali Ecology, continuously enhances the characteristics of data analysis and lowers the threshold of analysis. In the past year, this service has been widely applied in 10+BU business groups such as wealth, micro loan, insurance and group customer experience, greatly improving the efficiency and completeness of analysis. At the same time, many business lines quickly build vertical data applications based on ant data analysis platform, which greatly improves the construction efficiency by 50%. In response to such rich data analysis scenarios, how to build a data analysis platform to support these scenarios has become the core challenge of our technology. In this share, I will start from a business scenario, and work with you to deduce how to turn business problems into technical problems, and how to solve these technical problems.
[Guest introduction: Zhang Ronghua, Senior technical expert of Ant Financial]
I am currently in charge of ant Financial’s data analysis platform, including business intelligence analysis products and scientific computing related products. I have worked in the Internet field for many years, and have systematic understanding and precipitation in computer science and technology, abstraction and architecture, and data analysis.
4. GeaBase and its application in ants
Time and Introduction: 2018/5/10 15:50-16:30
As the application of big data becomes more and more widespread, people gradually begin to pay attention to the internal relationship between massive data. How to quickly discover valuable information in a huge network of relationships will be a big challenge in the field of big data in the coming years. Gea Base is a new generation of financial real-time graph database developed by Ant Financial Completely independently. The supported ant Financial risk control network, capital relationship network and other key application scenarios have reached 10 billion nodes and 100 billion edges of massive data scale, ensuring the risk control of payment, anti-money laundering, anti-cash, financial case trial, friend recommendation, financial information recommendation and many other businesses. In this sharing, we will introduce the graph database and graph computing platform, and its application in ant business.
[Guest introduction: Fu Zhisong Ant Financial Graphics computing and storage technology team senior expert
He joined Ant Financial in 2015 and is now a senior expert in Ant Financial’s graph computing and storage technology team. Participated in the development of real-time graph database GeaBase as the main architect and researcher, and realized the complex query and change of super-large scale relational network at millisecond level. At present, the team’s main responsibility is to develop a new generation of graph database and graph computing products. He received his B.S. in computer science from Zhejiang University and his Ph.D. from the University of Utah. After joining BlazeGraph, he led the design and development of distributed graph computing system.
For those who want to absorb more technical dry goods, come to the DTCC 2018 conference and listen in person!
— the END —