Academician Ni Guangnan tells you | 'China Core' | Why is it so difficult?

Source: National Financial Weekly

In the network information industry, the chip is likened to the heart, and the operating system is likened to the soul. In the past 20 years, why has China's Internet Industry been 'core-less'?

In April 2018, the U.S. Department of Commerce banned ZTE from purchasing U.S. chips. ZTE was 'stuck in the throat'. This revealed China's lack of independent innovation capability in the high-end chip industry.

For the missing issue of 'China Core', Caiguo Jun interviewed Ni Guangnan, a 79-year-old academician of the Chinese Academy of Engineering.

Interview guests: Ni Guangnan

An academician of the Chinese Academy of Engineering, graduated from the Radio Department of Nanjing Institute of Technology (now Southeast University) in 1961, and worked in the Institute of Computing, Chinese Academy of Sciences. He has been engaged in the research and development of computers and their applications. He was involved in the development of the first electronic tube computer designed by China. For many years, he was committed to promoting open source software such as Linux in China, promoting domestic chips and operating systems, and providing suggestions for promoting the formulation of relevant industrial policies.

In the pain of the ZTE Incident, we must hold the core technology!

The core technology is unattainable

The general secretary’s series of Internet-based letter speeches actually stated that, for example, when it comes to the mastery of core technology, it is to solve it by independent innovation. We cannot just hope that people will help you, or sometimes that we will buy (technology) ) It's OK.

At the Institute of Computing Technology at the Chinese Academy of Sciences, we are big machines. Historically we have been working as a large-scale computer. What we feel most deeply is that large-scale computer users have been blocked and cannot be sold. Sometimes there are places such as seismic exploration and weather forecasting to buy large machines. , At that time, the foreign company wanted to send someone to look at you in the computer room and see that you used it in this place. It should not be used in other places.

The development of our high-performance computers is relatively good. There are 'lights of the Shenwei Taihu Lake' and 'Tianhe'. Now the world is the 'light of the Taihu Lake' and the second is the 'Tianhe'. These machines compare our development in the world. Well, because these robots don't sell to us, they only make their own innovations.

Important Internet products must be created

Some products are worse, they are easier to get from the market. Doing research and development from scratch and turning into their own products have a longer cycle and more investment. Therefore, many units are relaxed. In this regard, Huawei has done a better job. Huawei This aspect is very important.

Ren Zhengfei said I need to have a backup system. The operating system is more difficult. He said that we may not have the same operating system as others nowadays. But I have to have a backup system. If someone doesn’t give you a job, at least you can maintain it. So I I feel that this concept is very important. It is necessary to realize the importance of cybertrust security. We must make up our minds that important network messaging technology products must be owned by our country. We cannot expect that you can buy it under any circumstances, or that others can give it. Authorization, I think it may be possible to learn this lesson in the ZTE event.

Information security issues are also 'time bombs'

ZTE has exposed a problem, because you are not your own core technology, so you have to rely on people, and the supply chain gives people a neck.

However, I think the ZTE incident has not revealed yet another very important risk. It is the problem of cyber security and information security. Now many incidents, such as Ukrainian power outages, are completely paralyzed after being attacked. So if core technologies are not mastered, we may Being subjected to the supply chain to give people a 'knuckle' and the safety of being 'knocked', it actually caused a prism-like incident, or a power outage in Ukraine, or the Iranian nuclear facilities were attacked to undermine this situation.

After the ZTE incident, it should be deployed early and resolutely break through core technologies. This is a very important issue. There can be no luck.

After spending hundreds of billions of years to forge the 'China Core'

Can't say that Chinese chips can't

It is not that the ZTE incident shows that our chip is not enough. The situation in different fields is not the same. I just mentioned the field of high-performance computers. Our chips are not inferior to others. We can compare the chips on mobile phones with those of foreign countries. It may be that some chips, such as FPGAs, are indeed a big gap between us and foreign countries. However, after the occurrence of the ZTE event, it was generally said that Chinese chips would not work, and sometimes they did not meet the objective conditions.

Chinese chip short board is made

Actually, we think the chip can be divided into two major parts. Part of it belongs to the design aspect and part of the manufacturing. The two are viewed from the industrial perspective as the upstream and downstream of the industrial chain. China's biggest short board is manufacturing.

For the entire IC manufacturing industry, I believe that for a relatively long time, we must invest a lot of money to support it. Now that the state has a large fund, IC funds, and social funds have also joined in. It will take several hundred billion yuan to be effective at least. To continue to vote, I don't know how much money will eventually end up. It really takes hundreds of billions of them to be visible.

A production line is at least two to three billion renminbi, not billions of renminbi, and even one line can't do it. And this is the equipment for people to buy. If you want to put the industry chain into it, equipment self-sufficiency is more difficult. But the current practice is To invest money in foreign equipment, there are also a lot of experienced experts who have to break through some of the problems of the process, as well as the materials. So the industry chain is very long. From materials, processes, manufacturing, production, testing, packaging to the market, we know that it is really big. Investment requires a lot of money, talent input.

Made in China 2025, general industry is about the same as the second echelon about 2025, and the first echelon in 2035. This goal is achievable. I think that IC manufacturing may be close to traditional industries. Internet e-commerce can be very fast, there are people It is relatively easy to say that China’s Internet is now developing and running in some areas of the world. Integrated circuit manufacturing obviously requires a lot of equipment, capital investment, and of course, human intellectual investment.

Localization of chips cannot be driven by BAT

At present, BAT has not invested much power in core technologies such as chips. They have done well in such areas as e-commerce, mobile payment, and shared bicycles. However, in the traditional computer field, the CPU is also good, and the operating system is also good. They don't invest much.

It's not that he can't do it, but he hasn't been able to use it as the main direction at least until now. It's not about investing in this area. It's still relying on the country to do more work. It's like satellite navigation system countries do it because The field may be profitable for more than a decade. Many companies and venture capitalists feel that the risk is too high. The average company will not do this.

Industry competition is not like foes

Microsoft's 'Piracy' Strategy

In the Chinese market, piracy is basically something that vendors deliberately do. That is, Microsoft intentionally gives you piracy. When you install a software certification ID, every pirated software you install, like before you started from win7, is Microsoft allows you to install pirated versions. And piracy. The version is the same ID, and Microsoft knows who installed the piracy and who did not install piracy.

He asked you to install pirated software. Why? Domestically, it doesn't work, because my software is here anyway. He would rather not charge you to use this software, so this is strictly speaking not a piracy. It is a Microsoft strategy.

Internet companies never say that you are pirated because their charges are not in the software itself. Microsoft strictly said that it is unreasonable. You obviously intentionally used this software for Chinese users free of charge, so that Chinese people did not have the opportunity to develop their own operations. System, Conversely speaking, Chinese piracy, said Chinese thief. Our internet company APP is free for users, no internet company said that Chinese users are thief.

China Netcom Industry Lack of Cooperation

In some of my speeches, I commented on Chinese technology people. Our ability to innovate is relatively good. All aspects should be very diligent. But if we have weaknesses, such as team spirit, the spirit of cooperation is not very good. There are traditional cultural influences in this area. It's better not to make chicken tails.

Let me give an example. Several years ago, the four ministries and commissions of the country jointly issued a document. All computers must be pre-installed with operating systems when they are sold. We hope to give domestic operating system manufacturers a chance.

But because we were not united, some people started to say that they were pre-installed in hundreds of pieces, and they were competing in the next 10 or 5 pieces. Finally, as far as I know, all of them are zero (yuan), mainly because they can expand their influence. Very good, in fact, our company did not benefit.

To concentrate on doing big things

What is the most important advantage for our country? To concentrate on doing big things, we are more suitable for the development of monopolistic fields. But in reality we are not doing well now. For example, on our desktop, wintel is a, Microsoft. The operation is dominated by the world. Although everyone is based on the open source Linux development of the desktop operating system, you can at least see that there are still 78 or so, is essentially based on the same open source software developed, little difference, But the brand is not the same, the specific details, such as hardware and software support, the problem of adaptation is not the same, there must be no benefit, the technology can not develop, because you are distracted.

Your opponent has been unified. There is only one in the market. Our market is very small, the size of the company is small, and there are seven or eight companies to do. It is the responsibility of the state to integrate these decentralized resources into a larger enterprise. With a unified standard and foreign multinational companies competing, this will have the possibility of success.

20 years of operating system development road, but also how long?

From unavailable to available

From a technical point of view, software has a development process. Software version 1.0 may not be used at all, and it will be 3.0. For example, after Microsoft windows 3.1, it will be easier to use. Now, overall, domestic software is generally in the 2.0 phase. , is the available stage. Not available to usable, but it may be a little distance to use.

From the overall situation of the entire market, the technical system, for example, wintel, a typical traditional computer system, wintel is easy to use, the user feels good, and the cost-effective experience is good. Domestically, we feel that Can be used, but maybe the price/performance ratio is not good enough. The main reason is that the user experience will be poor. There may be applications that you cannot adapt to. This is possible.

We began to try to replace wintel. At the earliest, the Beijing Municipal Science and Technology Commission in Pinggu County in 2005 first tried to replace wintel with a computer desktop system made of hardware and software.

At that time, I felt that I was still in an unusable stage and I was not in a usable stage. So it was very hard. Someone would put two computers. The leader came to show him that I was using a domestic system. The leader left and used wintel. This is quite early.

Now I think it is available. I just said that the Aerospace Science and Industry Group's 'Commercial Network' has 40,000 units, of which 28,000 use domestically produced hardware and software. At first glance, there is really no wintel on the edge. All the work is done on it. It is also possible. Did it very smoothly. From 2005 to now more than a decade, to the usable stage.

How long does it take to be usable?

It's easy to use. I don't think that all the units will be reached at once. For example, we can use it in some official document systems. It's relatively simple and it can be very fast. Now, some places feel good to use, because he is these jobs. We are all doing well. That's it.

However, in some places, such as our design, some large-scale software for industrial design does not exist. You think it may be a little bit worse and it may take longer. However, as a whole, I think that the development is going from being usable to developing. It may be basically several years later. Easy to use, it may have a process.

Insist on research and development for many years, still need a lot of investment to do well

The big things require a lot of manpower. Huawei's R&D personnel are more than 80,000. Therefore, it is necessary for some relatively powerful companies to participate in R&D. Small enterprises can do some innovation, but the system is established, the ecological establishment, and those big ones Businesses must play a greater role.

You look at our Internet services. For example, e-commerce has the ecology of e-commerce, social networking has the ecology of social networks, search has the ecology of search, and big companies take it up. But like this universal machine ecology, desktop, mobile, The operating system is more difficult. Now it needs the drive of some big companies.

It is not the money that breaks the monopoly deficit. What is missing is the early heart.

National Science and Technology Development Plan

The country has this technology development plan from the top level design, and the special item gives money, but we are not good at execution, we have to reflect on it.

At that time, we established this major project to replace the foreign system. At that time, we may have tried our best to hide our backs. We haven't made it clear yet. The gap is also relatively large. But since the development, we have discovered that some of the 'nuclear high base' major projects do not follow the direction. Some projects support wintel. Because the general secretary asked for accelerating domestic self-controllable alternative plans to build a safe and controllable information technology system in his online letter work speech. One of the alternatives mentioned here is the system, the alternative is that you To use domestically produced alternatives to foreign ones, because people have monopolized, it is not a blank. We are not working on white paper, but on existing foreign monopoly of our market.

Unlike artificial intelligence, we now have a lot of people doing artificial intelligence chips. We basically started at about the same time as foreign countries. I dare not say who succeeds unsuccessfully. This is a new area, a blank piece of paper, but we are now facing Like the desktop, like mobile, people have monopolized, so you must have a process to replace people, to replace the monopoly, so this is more difficult.

Breaking monopoly needs system support

It is very important to form a technical system in the network message. One single product you say is good, for example CPU you do, the CPU does not have software, the CPU will not move, the CPU needs an operating system, the operating system has database software, middleware, each A variety of application systems, the use of hardware also requires some support, this is a complete system.

The general secretary of the Internet letter work speech is the most important top-level design. To build my own system, it is our very good direction. We must not forget our minds, and implement the country’s original plans well. Don’t change. Like this.

2016 GoodChinaBrand | ICP: 12011751 | China Exports