Sci-tech creator, Digital China Group CIO Shen Yang: The best practice model is failing, open source accelerates distributed innovation

Shen Yang, Vice President and CIO of Digital China Group
He holds a bachelor's degree in engineering from National University of Defense Technology and a master's degree in engineering from Shanghai Jiaotong University. He used to be the consultant of SAP China Global Support Center, the technical architect of the SAP US digital transformation service department, the head of the enterprise planning and business intelligence team, and the chief information officer of Digital China Holdings Co., Ltd.

—
Text | babayage
Edit | Laugh Laugh

In the communication with Shen Yang, I felt that he has a unique "alienation" temperament: not obsessed, not fanatical, and constantly examining the causes behind the direction of life from a third-party perspective during the growth process. Compared with the introverted and enthusiastic idealist, he is more like a rational and calm observer who sees the world as well as himself.

Understand the meaning of information from your growth experience Kechuang people: When you make major decisions in life, do you have a consistent principle or an ideal you have been pursuing?

Shen Yang: Judging from the trajectory of life, there are many choices that are not clearly thought through, and many changes are random.
As a post-80s generation, there was still a lack of information in the process of growing up. At that time, computers and the Internet were not yet popularized, and information could only be passed on by word of mouth. At the time of the college entrance examination, most students, teachers and parents didn’t know how to fill in their volunteers. After carefully filling in the first and second batches of volunteers, I curiously added a pre-approved volunteer – the Master of Science and Technology of National Defense University: Automation, and was later admitted directly in the advance approval. During the undergraduate period, students had the opportunity to participate in major projects such as maglev, Beidou, and brain-computer interface, but I still hoped to get in touch with the new world at that time, and wanted to study abroad and get in touch with the commercialization system. However, the 9/11 incident in 2001 made it extremely difficult to obtain a student visa in the next few years, so I chose to go to Shanghai, the most developed city in China, to study for graduate school.
Joining SAP after graduation was also accidental. When I was a graduate student at Shanghai Jiaotong University, I was fortunate to do some work related to the testing and development of mobile display chips in the Intel R&D Center. As a result, Intel sold its XSCALE chip business at the end of 2006, and my original employment plan was disrupted. It happened that SAP was expanding around the world at that time, and it was a coincidence that it joined in. The first time I took a plane in my life was actually an international flight (laughs).
When I was at SAP, I made an active choice. At that time, most of my colleagues were more interested in the more business-oriented finance, sales and distribution modules. Of course, these are also the core competitiveness of SAP. However, it may be the reason for the lack of information in the process of growing up. I am more interested in data and information, so I chose the data warehouse and data reporting module team, and specialized in the most popular budget planning module (PLANNING), this module is through Planning to remove uncertainty from business operations. The general data report can be understood as addition and multiplication, while the PLANNING module is mainly for division, which belongs to the reverse engineering of addition and multiplication, and the threshold is relatively high. The disadvantage of the unpopular module is that it often sits on the bench at the beginning, and there are not as many opportunities to do projects as other colleagues; but the advantage is that at the beginning, there is enough time to follow the tutor and learn more deeply. Later, when the business of this module became popular, the company had to call consultants in this field all over the world.
After I went to work at SAP in the United States in 2010, I continued to work in the big field of data warehouse, reporting and budgeting, and also experienced the initial implementation of the SAP HANA in-memory database. Later, SAP merged the products related to enterprise planning in the enterprise into an EPM module (Enterprise Performance Management) product, covering the financial consolidation and various business budgets of the enterprise. Later, there were more and more customers for this product. Most of the new customers were in the United States, and most of the new development teams were in China. I also cooperated well with the Chinese development team in the United States, so I began to develop and develop in the United States. lead the team.
Looking back, I do have a lot of randomness in my professional experience. When I first started my career, I was still very anxious about this randomness. Fortunately, the general direction of data and information is also my interest. However, it was deeply influenced by the philosophy of quantum theory. After all, the world is not deterministic, and people have to accept probability and randomness.

Kechuang people: Just a small question, have you ever thought about returning to China to start an EPM product?
Shen Yang: At that time, most of the customers for EPM products were still in the United States, so I really didn't think about going back to China to do this. But my previous SAP colleagues have done it, and they have done it well, I think I may not have a chance (laughs).
Many random and accidental factors continue to affect my growth process, but the things I have been paying attention to in this process are the absolute amount of information, the cost of acquiring information, and the speed of information transmission.
I liked reading encyclopedias when I was a child, and started reading physics textbooks in high school and university when I was in junior high school. At the beginning, it was not because I liked it, but in the non-Internet era, there were not many books around me. Can only see what. This is a bit like a full table scan in a small database, where the amount of information that can be obtained is limited and the cost of obtaining information is high. If there was a large library nearby at that time, and there were good book lists, catalogue guides, and expert recommendations, you might have discovered your interests and expertise earlier.

Kechuangren: If you had the opportunity to go back in time and send a message to your past self, what would you like it to be?
Shen Yang: (thinking for a moment) Combined with my own growth experience, I don't think this imagination is so attractive (laughs). I think cognition, knowledge, and high-value information are all resources in the digital world, and human growth depends on how well the digital world matches the resources of the physical world. , Second, there is no application scenario, and this knowledge is not so meaningful.
If you really want to send a message, I hope to tell the past me not to be so anxious about randomness. However, this doesn't seem to be a piece of information, and you have to experience it yourself to understand it.
In the past, I would also imagine what it would be like to have a panoramic plan of life very early. This idea was cut off because I realized the key to the matching degree of information and material. If there is no suitable guidance and matching resources, it may not be able to go on. Instead, it may be better to adjust and adapt according to local conditions and situation.
Life growth and digital transformation have many principles that are interlinked. In the process of digital transformation, it is not that the more data is collected, the better. Ultimately, the system must be transformed based on resources and business goals. There is a joke that a soap box production factory wants to detect which soap boxes are empty. A computer doctor team tossed AI recognition and automation for a month. As a result, a worker brought in an electric fan and blew the empty boxes away. Personal growth is to find the right fan at the right time, and piling up a lot of information sometimes becomes a constraint.
Excess profits only depend on the moat and the data chimneys will only grow

Kechuangren: What are the details that impressed you the most during SAP?

Shen Yang: When I first joined SAP, most of the colleagues who joined together were not computer majors, and there were even a lot of pure liberal arts backgrounds. Everyone spent more than half a year receiving various technical training and training on projects, including business etiquette. And transnational culture, etc., and even learn a driver's license in a foreign country. This diverse and international atmosphere enables us to deal with a variety of complex issues. When we interview students who majored in physics, we can find many employees who have graduated with a doctorate in physics to discuss quantum mechanics with candidates; when communicating with customers in the chemical industry, experts can also use details to gain professional recognition from customers. However, domestic companies that make cross-industry application software and solutions rarely have the resources and patience to invest in such training.
Kechuang people: Can these complex training systems, especially business etiquette, really bring more profits to enterprises? If not, why are foreign companies willing to invest resources to do it? If so, why don't domestic companies do it?
Shen Yang: This kind of decency and professionalism can create value, but it is a top part of the value pyramid and needs a solid profit base for the company. There is only one source of true excess profit: the moat. "Can create value for customers" is the pass line, "Only you can create value for customers" is the high value line. Enterprise strategic decision-makers should constantly think about and invest resources to build a moat. Substitutable products and services are not enough to ensure the company's profits, let alone excess profits.
First, some software companies in China have a relatively short creation time, and they also need to wait for the customer's concept to change. There is not enough accumulation to create their own value system and moat. I believe that when they develop to a mature stage, companies will definitely build a complete training system. .

Kechuang people: The Digital China Cloud Base project you have managed has achieved good revenue by exporting Digital China's digital transformation capabilities to the outside world. What is your company's moat in terms of digital transformation?

Shen Yang: I think the moat of the cloud base is the ability to combine enterprise service experience with IT technology and ecology, as well as a deep understanding of large-scale enterprise application scenarios. For example, we have deep cooperation with the community of the open source software TiDB. Most of TiDB's initial customers are Internet companies. The business of Internet companies is actually very flexible. For example, when tens of millions/hundreds of millions of users are encountered all at once, the amount of data expands extremely large in an instant. Such users pursue elasticity, scalability, and high availability. In addition, TiDB is compatible with the MYSQL protocol, making it easy for many business systems to migrate to the new database platform in the past.
Internet companies use MYSQL a lot, because there are many massive data scenarios on the Internet, and the data chimney barriers are not high, and many times a large table can be used in the world. However, in large non-Internet companies, there are naturally many chimneys, business chimneys, and data chimneys; the user elasticity of internal applications is not so high, and an enterprise's employees will not suddenly increase from 10,000 to 100,000. Therefore, in the internal application of the enterprise, complete functions, multi-table association, and high availability are the more important demands of the database. In this regard, generally speaking, the PostgreSQL protocol database performs better than the MYSQL protocol database, and many of our internal enterprise-level applications are built on the PostgreSQL database. In the US, PostgreSQL is about half as active as MySQL, but in China, it's probably less than one-tenth as active.
There were CockroachDB and YugabyteDB based on the PostgreSQL protocol in the United States, but there were no similar products in China at that time. Therefore, based on the TiDB product, our team hopes to add adaptation to the PostgreSQL protocol to fill the gap in this direction, so that a database can support both MYSQL and PostgreSQL protocols.
On the other hand, we also hope to use this architecture to support our internal application system, and hope to use a distributed database and a low-code application platform to complete the construction of most agile applications.
Of course, the difficulty of this project is also relatively high, and we need to rely on the strength of more communities.

Kechuang people: Eliminating data chimneys and data islands is the slogan of many digital products, but do you think this is impossible?

Shen Yang: The chimney is not caused by everyone deliberately, but because of legal and regulatory factors. For example, the law stipulates that the financial data of a company cannot be disclosed in advance. Before the financial report is released, only a few employees know it. When the financial report is released, it must be disclosed to everyone, and not all detailed data is disclosed. The same is true of human resources data. It is impossible for everyone to know the salary data of the company. The affairs in the enterprise are operated by professional departments.
You may think that the data chimney is caused by technology. This may be the case when an enterprise uses a large number of commercial software. Not every commercial software can provide a complete interface to the outside world. The current data chimney is more caused by permissions. Even if all the data is in the same database, the enterprise will add various permissions to the use of the data. The data complexity of the enterprise lies in the association between permissions and multiple data tables. .
Data desensitization, data association, data permissions, privacy computing and security have become a hot topic now.
Reasons why the open source model will be successful:
Best practices fail, distributed innovation rises

Kechuang people: In addition to the cooperation with TiDB, you have expressed your optimism about the development potential of the open source model many times. Can you share your value perception of the open source model?

Shen Yang: The complete expression is: as the "best practice" model fails in more and more industries and fields, distributed innovation will become the mainstream, and the open source model can efficiently support distributed innovation.
The failure of "best practice" is an important premise. More than ten years ago, everyone bought software and liked the complete solution of the giant. One factory and one supply chain polished the operation system to the best, and everyone used it directly for reference, which is the most efficient. But today, a benchmark in the industry took several years to polish the best practices, and the external technology and competition system may have changed again. After directly copying and going online, it may be a generation behind.
Taking CRM as an example, traditional CRM has become SCRM in the era of mobile social networking. In the past, maintaining customers was maintaining a mobile phone number. A customer feedback demand required filling out a work order and going through the process. Today, everything may be a matter of pulling a group. This set of The system has changed the way of acquiring and maintaining customers.
And this is not the end, the changes in this era are frequent and drastic.

Kechuang people: Do you tend to think that "the new stable state replaces the old stable state", or "the best practice model of this stable state will be replaced by continuous instability"?

Shen Yang: I believe it is the latter, and it should continue to change for some time to come. Therefore, the value of open source is reflected. The times are changing too fast, but the iteration of the traditional software development model is not so fast. Even in the SaaS model, collecting requirements and optimizing iterations is a centralized process. There are inevitably time differences and trade-offs. Open source, a bottom-up, instant, distributed innovation model, can better solve problems. For example, if we encounter a pain point, I have the ability to solve it. I may open source it on the same day, and industry partners can use it on the second and third day.
This cycle is much faster than the traditional development model and business model. In the past, if there was a small pain point, the commercial software went through the procurement process, and the problem that the process may not be completed may have disappeared. However, developing and solving problems in the open source model requires a very short period of time. As long as it is effective, the more mature enterprise-level customers will have enough willingness to pay.
There are many advantages of the open source model, such as being more professional, commercial software often cannot achieve the ultimate segmentation, because there are too few dishes to fill a table, no matter how small the sparrow is, it must have all the internal organs, but open source software can be extremely focused; In terms of the time cost of judgment is very low, there are basically one or two left in a very subdivided open source field.

Kechuang people: Can you share Digital China's operating experience for the open source community?

Shen Yang: We recognize the importance of the community. Open source is not only the open source of code, but also the open source of the knowledge organization system and the interaction of the community. In terms of open source, we are just starting out, and we are in the process of continuous learning.

Kechuangren: You mentioned at the beginning of the interview that you are very concerned about the speed of information flow. Although everyone shouts that "the world's martial arts can only be fast and unbreakable", you seem to be more radical, thinking that saving the procurement process and selection time is all about Is it of great value?

Shen Yang: All thinking must revolve around the essence of the world, matter and time, and time is the most scarce wealth. Open source and information transparency are both conducive to simplifying the customer's procurement process and saving time for model selection, saving customers' decision-making costs and saving product sales costs.
It is also based on this thinking principle that I am also very optimistic about the new generation of low-code mode. Some friends think that low-code is just drag and drop, no different from the products of more than ten years ago. But in my opinion, the future cloud-based low-code is an economic combination of API + ecology. In the future, when you are doing development, you will first compare the cost of my own development in the application module with the cost of calling other people's services. The more developers will complete innovation based on a mature ecosystem.

The prediction of the future development of the open source model is optimistic about low-code, and customization is still just needed in the short term

Kechuang people: What judgments and expectations do you have for the future development of the open source model in China?

Shen Yang: First of all, there will be more professional open source companies, and there will be many new open source ecological participants like Digital China. On the one hand, we can directly contribute to open source, and on the other hand, we can verify open source products through our own scene experience, adapt and connect different open source products, and quickly fill in some blank areas, which in itself can bring great benefits to our peers. s help.
Second, the SaaS model will enter a tipping point. Many SAASs are still competing with each other, shielding interfaces from each other. China's SaaS industry will evolve into an ecosystem of several large platforms. More and more SaaS applications will grow on the platform. The underlying data and interfaces will continue to provide nutrients for the application ecosystem, and they will integrate and support each other. A good SaaS will drive another A good SaaS, in turn, drives another good service.
Third, as mentioned above, low-code based on cloud + API will gradually grow up.
Fourth, customization needs will still be Chinese characteristics. Compared with foreign countries, the proportion of software customization in China is very high. Even in the same industry, different companies have very different management needs.
This difference may be related to the current corporate organizational structure in the country. Foreign companies have been developing for a long time, and many of the management of most companies are also professional managers who have received MBA training. The unique management style is easily diluted.
In addition, many industries have very strong industry standard systems, and companies in the same industry have little difference in management behavior, so SaaS solutions in many industries have been born.
Comparatively speaking, the development history of modern enterprises in China is not long, and most of them have very unique management styles. Coupled with the rapid economic development in China in the past few decades, there are endless opportunities in all walks of life, and enterprises are constantly exploring new opportunities and tracks. This has led to the situation that in the process of enterprise management, general solutions often fail to meet enterprise requirements.