Video | Binary Tree Team
Text | Cai Fangfang
He is one of the earliest practitioners in China to start promoting the concept, technology and business of cloud computing; he has won the title of "cloud computing brother" in Microsoft China, Microsoft China's first private cloud solution, the first cloud computing demonstration center, the first private enterprise customer case, the first cloud computing white paper are all from him; he is the first technical staff of aws China, and later led the team to build Ping An Cloud from 0 to 1, also known as the "Father of Ping An Cloud"; he called himself the public cloud "dead multi", thinking that the future belongs to the public cloud. Private clouds are just a compromise in evolution. At the Ping An Technology Building in Shenzhen, we interviewed Fang Guowei, a guest in this issue, and listened to him talk about the decade in which cloud computing has subverted traditional IT.
The guest of this issue of Binary Tree video is Ping An Technology CTO and Chief Architect, Fang Guowei.
William Fang, CTO and Chief Architect of Ping An Technology. Graduated from the Department of Computer Science of Tsinghua University with a master's degree. He has served as the director of the Huawei Central Software Institute Erection Department, the chief cloud technology expert of Amazon aws China, and the senior architect of the cloud computing "Center of Excellence" (COE) team of Microsoft's global services division. He has edited "Making the Cloud Accessible" and "Detailed Explanation of Microsoft Windows Azure Cloud Computing Platform", and has delivered many keynote speeches on cloud computing at domestic and foreign technical conferences.
infoq: You were called "Cloud Computing Brother" by your colleagues when you were at Microsoft, can you tell us how this title came about?
Fang Guowei: Microsoft Global released a cloud at that time, but it was difficult to land in China, at that time I felt that there was a problem with this cloud method, I told the leader, we want to build a set of private cloud solutions to help Microsoft achieve revenue in the business, so that on the one hand, it can help Microsoft China establish the influence of the cloud, on the other hand, there can be some actual cloud revenue. The company thought it was a good idea, and it eventually became our overall strategy, and I became the owner of Microsoft China Cloud. Just having a owner is not enough, because the cloud thing is new to most of the employees of Microsoft China, and I have to learn, understand, and design the plan myself, and at the same time, I want to give everyone literacy. At that time, the concept of cloud computing was very new, and most people did not understand the cloud. At a conference in 2008 and 2009, the moderator asked the people below what cloud computing was, asked 10 people to get 12-13 answers, and some people gave more than one answer.
I will tell people inside the company what cloud computing is, what is the benefit of doing this thing for Microsoft, why cloud computing is the future, and slowly subverting traditional IT. Coupled with the fact that I took the lead in writing Microsoft China's first cloud computing white paper, and then did the first chinese cloud customer to land, and wrote a book about Microsoft cloud computing, so everyone thinks that I created cloud-related things in the company.
At that time, at a departmental review meeting, the department leader said that "Fang Guowei's cloud thing is doing a good job", and at that time, someone said that "Fang Guowei is the father of Microsoft cloud computing". The leader said that it is not okay to call the father of cloud computing, he is still very young, just call him a brother. Since then, everyone has called me Microsoft China Cloud Computing Brother when they meet, and slowly this title has spread.
infoq: After microsoft, did anyone call you that?
Fang Guowei: There are fewer people outside, but many people inside Microsoft know that even after I left Microsoft for many years, they would joke that this is Microsoft's Chinese cloud computing brother. In September last year, we received a former Microsoft colleague from Hunan Financial Holdings, who brought the vice president of their group to Ping An to visit, and at the meeting, he also said that this was our cloud computing brother at that time.
infoq: So this title has always been with you.
Fang Guowei: Yes, many of Microsoft's former colleagues, including many employees of Microsoft China, know. Because I did a lot of cloud computing-related work, as long as there was a cloud problem at that time, everyone knew to find Fang Guowei, and slowly established his own reputation. I also say this to my colleagues in the current department, you have to do anything, especially to do technology, first of all, you must set a goal, do things to do the best in the team, after the team does the best, then do the best in the department, and then do the best in the company. If you are strong enough and have a certain reputation in the industry, you can establish your own reputation. Just like the product must have a good reputation, everyone's career development is the same reason. A person's knowledge system architecture, the most like the letter t, has a certain width, but in a certain field to do a needle to break the sky, so it is easy to reflect the value, but also let your leadership can not be separated from you, become the leader's right hand, I think this is what many people need to do in their daily work. Of course, not only in the field of technology, but also in other daily jobs, only after becoming the right hand of the leader, will you be successful in career development. Here's an interesting metaphor, suppose the leader has a meeting with the team, the house is on fire, the leader can only take away two people, you have to think about whether you are in these two people, if you are inside, then congratulations, you are in the leader's mind is the right hand, your position in the team is no problem, you don't have to worry about layoffs.
infoq: At that time in 2008-2009, the industry's definition of the concept of cloud computing was actually very vague, and there were all kinds of definitions, how did you learn and sort out this whole system at that time?
Fang Guowei: Indeed, at that time, the concept of the cloud was controversial, some people said that the old wine was in a new bottle, and some people said that this was a change. To sort out this matter, we must first pursue the face, and all kinds of statements must be understood and seen. I was asked to read at least one technical paper a week, and then I read it too. There's a book called "It Doesn't Matter Anymore," and that book is very good, from Edison to Google, about the big shift in it, and the commanding heights are cloud computing. Of course, you can't just absorb information, but also have your own thinking, you have to think about why some people think that the cloud is the development direction of the whole IT, how you think about yourself, according to the customers you contact, your own daily work to think, this statement is not reliable. When I approached the cloud in 2008, I had been working for about 8 years, and before joining Microsoft, I worked at IBM, where I had been in contact with many customers and had a certain understanding of the application areas of IT. After being exposed to the concept of the cloud, I felt that this matter was reliable and a direction. The most important feature of cloud computing, one is the economic scale, the second is automation, and the third is the professional division of labor, these three things are determined from an economic point of view, it is indeed better than the traditional model, can replace the traditional model. So I made up my mind that my career development should also rely on this aspect. If you want to go against the trend, it is not necessary to say that it cannot be done, but the difficulty will be much higher, if you go with the trend, many things will have the effect of doing more with less.
infoq: So at that time, you decided to do cloud computing as a kind of homeopathy?
Fang Guowei: Yes, I thought so, and at that time, I also believed in my heart that this matter was indeed the direction of the future. Personality-wise, I don't really like doing repetitive work, and I don't think it's much fun to do repetitive work, so I'll try to do something new, and the cloud was really new at the time. Coupled with the fact that in the past 10 years of development, we have seen an endless stream of new things in this field, proving that this judgment is not problematic, and it has indeed slowly succeeded. If in the last 10 years, you find that it is not very good, you can't go all the way to black, but you have to adjust the direction. In the past 10 years, I have been relatively lucky, and the judgment at that time was correct, the form of the cloud may be different, the work unit has changed, but the field of my focus has not changed, and I have always done cloud computing.
This is a bit like a principle mentioned in "The Alien", which is the ten thousand hour theory. I didn't work on a cloud job for 10 years, but I tried different roles. Public cloud, private cloud, hybrid cloud, private cloud, etc. all kinds of clouds have contact, from iaas to paas, saas, etc., have done Party A, but also done Party B, the understanding of the cloud is relatively comprehensive.
infoq: You think the cloud is the big trend, why did many people think that cloud computing was "old wine in a new bottle" at that time, and there was nothing new?
Fang Guowei: The same thing, different people may look at different angles. People who think that new bottles of old wine may be concerned at a certain point, cloud computing many technologies are not produced out of thin air, but equivalent to a continuation of the previous, but there are innovations in the process of continuation, and there are great changes in business models. Why did I think this thing was reliable and could be done? A big reason is not just technical, but more importantly, its business model has changed a lot. For example, in general, companies will procure their own servers, recruit programmers, system administrators, and build their own IT systems. This method is more decentralized, in fact, it is a manual workshop method, which is very inefficient. It's just that it's in the IT field, and it feels like it's very high-tech, but in fact this service model is very rudimentary. Cloud computing is a large-scale production, and I think this large-scale production model will definitely win the manual workshop method. I think the essence of cloud computing is the servitization of scale and automation in the IT field. There are three key words in this sentence, the first is scale, the second is automation, and the third is the embodiment of service. We often say that cloud computing is equivalent to turning the fixed assets of traditional IT into on-demand use, which is the embodiment of servitization, which is also the essence of what I think of as the cloud. From this point of view, I think cloud computing is easier to understand, it is determined by the economic basis. As for what kind of technology is used and what kind of virtualization is used, it is the means of implementation. Virtualization is not the use of containers, with what kind of stack, these are not the key to the problem, the implementation of the means can be different. Many people have a misunderstanding when they are exposed to cloud computing, especially those who have just been exposed to cloud computing. He will feel that using cloud computing is to choose a stack, such as openstack or cloudstack, and his focus is on many frameworks. In fact, after a period of exposure to cloud computing, you will find that the key to cloud computing is not the stack, the most important thing is what kind of technology platform is adopted at the bottom. Many cloud platform problems are not in the stack, but in the network, storage problems, resulting in large-scale failures, which is the most technically challenging place to do cloud computing.
infoq: After joining Ping An, you led the team to build Ping An Cloud from 0 to 1, what was the biggest challenge in doing this?
Fang Guowei: This challenge can be seen from several aspects, one is that when I arrived at Ping An, Ping An's team was relatively small, about 10 people, and at that time we called the innovation team, not the cloud team. Everyone understands the cloud relatively little, and I am the first person in this team to have a long period of contact with cloud computing and have really done cloud computing related products. The team needs a lot of changes in concept, including how the architecture should be designed and how the product should be designed. At that time, resources were also very limited, and it was difficult for a dozen people to make a cloud platform for Ping An, so at that time, I first proposed to "set up a stage and sing at Ping An". I bring the cloud platform team to help you build a platform, welcome other teams within the company to join, build the cloud platform of Ping An, and do this with a relatively open mind. When we did the cloud in 2014, the traditional technical architecture has been developing in Ping An for many years, and there are still many challenges to the cloud as a new thing, and other teams will feel that you are not robbing their business. So we proposed at that time to double mode, the new application of the cloud model, the old application of the old model, so slowly moving forward. Another challenge is cloud awareness and resource investment. The cloud itself is highly technically demanding, requires significant investment, and may not see an immediate return in a short period of time. Even the industry-leading Amazon AWS takes a long time to become profitable. We have to take the cloud as a business, we must take it very strategically and be willing to invest for a long time, which requires us to communicate with the top level repeatedly. Now that ping an problem has been slowly solved, on January 1 this year, Ping An changed the entire logo to "Ping An of China Finance and Technology", which is not like the previous support business, technology itself has become a business, and the dependence of technology on the cloud is very strong. Ma Zong also proposed that "those who get the clouds will get the world", the position of the cloud is very high, and after the strategic attention is paid, many things will be easier to do.
infoq: In the process of doing Ping An Cloud, did you encounter anything particularly difficult that made a special impression on you?
Fang Guowei: The thing that the pressure of doing the cloud is that it is a bit like a double-edged sword, on the one hand, you hope that the cloud will develop better and better, and more businesses and applications will be put on the cloud platform; on the other hand, when more and more business is put on the cloud platform, the stability and security challenges of the platform will be very large, and your pressure will be very large. For example, most of Ping An's internal business runs on the cloud platform, which has high requirements for the stability and security of the cloud platform itself. When hundreds of billions of large businesses are running on your platform, once there is a problem, it will affect the business, so the pressure is relatively large. Especially in 2016, when the technology was still in the process of polishing, the challenge was very big, and I was under a lot of pressure. I've been exercising, I'm doing well, and in 2016 I had a stomach ulcer because of the pressure.
infoq: Do you think doing cloud computing is hard?
Fang Guowei: Cloud computing is at the bottom, and it is indeed not easy. My WeChat signature is "The cloud road is long and its repair is far, we will go up and down and explore". Many company executives don't understand the technical challenges of doing the cloud itself, and its business effects aren't as immediate as many applications. This is the same as our country's investment in basic scientific research, the investment in applied science and technology is easy to see the effect, and the investment is not so large, but the investment in basic research and development is large, and the effect is slow. I think cloud computing is such an industry, especially the underlying cloud platform, all companies that want to make the cloud platform must attach great importance to it strategically, be willing to invest a lot, and have the patience to do this thing, which is not easy to do cloud-related practitioners.
infoq: You just said that the cloud industry is at the bottom, and to precipitate, it may invest a lot, but you can't see a lot of output. Did your involvement in the industry have anything to do with your personality?
Fang Guowei: People who do cloud really need to be calm, because it is slow to see results, it belongs to a service at the technical architecture level, unlike the application layer, it is easier for ordinary users to understand. You say you've made a very good technology at the bottom, and professionals may be able to understand what's great about the thing itself, or what's good about it, but it may be hard for the general public to get to. So only if you have a calm personality, like cloud computing, and sit on the cold bench for 10 years, it is possible to do things well. And the more cloud computing is done later, the more it needs the underlying technical research, and it takes a long time to accumulate to make some good innovations, and then reflect it in the final cloud product.
It should be said that there is a bit of a relationship, you have to be calm, endure loneliness, and be willing to persist for a long time, the pressure is still quite large. It's not that the pull is too long and your pressure becomes less. Once the cloud is online, the challenge is that there are many customer applications in use, and the work you have to do is a bit like changing engines while driving an airplane, and you have to do some upgrades and maintenance on the cloud platform, but the customer applications on the top cannot be stopped.
infoq: What kind of experience do you think you had as a child that made you develop such a calm personality?
Fang Guowei: I think there are several factors that have a certain impact and relevance on what I am doing now. One is the character of my father, my father is a very serious person, and he is willing to grit his teeth and go forward desperately when he does a lot of things, which will also have some subtle impact on me, so I generally do things more seriously, and the sense of responsibility is stronger. Frankly speaking, there are many bad things about a strong sense of responsibility, and the bad part is that you put a lot of things in your heart, especially the operation and maintenance things, and the pressure is very high. I suffered from a stomach ulcer in 2016, which also has a lot to do with my psychological stress. So there are good and bad, the good part is that you do things seriously, you can definitely do things well, and the bad part is that the pressure is relatively large. Another point is that when I was a child, my family had a big fish pond, and one of my big jobs was to go and see that fish pond. It's a boring thing, there's not much to do, but there's a lot of time to spend every day on vacation. I basically read there and allowed myself to endure loneliness. So I never feel lonely now, and I can often fill up my time and do something that enriches me, whether it's reading, studying, or exercising.
infoq: You mentioned earlier that you usually like to read, what books have had a great influence on you?
Fang Guowei: My philosophy is that if the body is healthy, it needs to exercise more, and if the brain is healthy, it must read more and learn more, and reading is equivalent to the movement of the brain. I was impressed by several books, one of which I read not long after graduation called "If I Knew Then What I Know Now." I would like to recommend this book to our new employees who have just started, especially those who have just graduated, and hope that they will have their mentality set. When I graduated at that time, I felt that I graduated from a good school and could do a lot of things, as if the company did not use me well. In fact, this view is very problematic, you should emphasize what kind of contribution you have made, what kind of performance you have made, and how the company is, not the other way around. Many times I have a bad mentality, I think hr should not take care of me a little, this view is wrong, should be like the book said "you are 100% responsible for whether you are happy or not", emphasizing that you have to be responsible for whether you are happy or not, rather than being helped by others to help you solve this problem. I think this is a very important point for newcomers who are just entering the workplace. I still read this book sometimes.
There is also a book that I read later, called "The Outlier", which I was also very impressed with, because this book explains a person's success better, in fact, there are many factors in it, not only do you have good skills, you are very smart, you iq is very high, in fact, you need a better platform, have better opportunities and so on to get together to succeed. It was a concept at the time, and I still remember it, that it said that a person should be successful, and the role of iq is like height and playing basketball. Generally, the higher the height, the better it is to play basketball, but in fact, after exceeding a certain height, the height is not so important. Just like the best person in the NBA who plays basketball is not necessarily the tallest person, it is not necessarily the most successful person in the current situation that iq is the highest. In our daily work, many times we overestimate the role of iq, and the role of other aspects is not so highly estimated. Another point is that the relationship between a person's role and the entire platform, just like I am doing cloud computing in Ping An now, I think the platform itself is still very important. Their own ability is on the one hand, and there are still many problems of opportunity. Why did cloud computing start to slow down in 2006, 2007, and 2008? The reason is that at that time, many of its technologies, whether it is virtualization or network technology, slowly matured, and got together to do things, which is the effect of platform accumulation, and the time point is also very important.
infoq: Are the developments in the domestic cloud computing industry consistent with your expectations ten years ago?
Fang Guowei: It was a little slower than I expected. The advantages of the cloud are obvious, there should be more people willing to use cloud computing, but from the perspective of the entire industry, the acceptance of the cloud is still relatively not as fast as I thought, so it takes more patience to do things. But in the long run, I think there is no problem. What's different from before is that there are still many people who are willing to choose private cloud and hybrid cloud, and I think it still takes time. The private cloud solution is a relatively short-term solution, a bit like ibm hosting. The use of ibm hosts has dropped significantly many years ago, but it is also used now, and it has not completely disappeared, but from the overall trend we all know that this is not a trend. I think the same is true for private clouds, and there will definitely be people who will use them in the short term. But in the long run, I think private clouds, like ibm hosts, will slowly die out or slowly decrease. Not only for economic reasons, but also for technical reasons. Public cloud innovation is very fast, and there will be a lot of products and innovations on a public cloud platform. But most private cloud products are only a small subset of the public cloud, and the private cloud is far from meeting the needs of the enterprise, at least it needs a hybrid cloud approach. So I'm optimistic about the public cloud as a whole, and there's no problem in the long run.
infoq: But now the private cloud is still quite hot, there are many companies doing it, including Ping An itself is also doing, what do you think is the significance of the existence of private cloud now?
Fang Guowei: I think a very important significance of the existence of private cloud is that it can digitize and modernize the existing technical architecture. It allows users to adopt cloud technology to do something in a relatively psychologically acceptable situation. I believe that in the end, we will gradually develop into public and private clouds through private cloud and hybrid cloud, which is a gradual evolutionary process. Many times it takes time for users to accept new things, including the establishment of trust relationships, and it takes time to prove, but I think it will slowly move to the direction of public and proprietary clouds.
infoq: So is it understandable that private clouds and hybrid clouds are now a compromise on the road to the public cloud?
Fang Guowei: That's right. I think that private and hybrid clouds are just an intermediate stage on the road to the public cloud, and may eventually move towards a state dominated by public clouds, with some private clouds and some private clouds.
infoq: When you were at Microsoft, you tried to promote the development of private clouds, which is not inconsistent with your repeated statements in the public speech later, "the future belongs to the public cloud"?
Fang Guowei: Yes, this is a stopgap measure, I have always thought that the private cloud solution is a step to the public cloud, many customers have difficulty accepting the public cloud at the beginning, he can start with the private cloud, and I believe that for large customers, he also has the demand for private cloud, which is also reasonable. However, in the long run, for the vast majority of customers, the public cloud is the direction, and the private cloud should be like the IBM host, someone uses, but there are fewer customers.
infoq: The theme of our issue of Binary Tree is ten years, hoping to stand on the timeline of Internet development and sort out several important time nodes in different fields. If you were to summarize the decade from 2009 to 2019, would you consider the past decade to be a decade in the cloud computing space?
Fang Guowei: Since 2009, the first 5 years have been more in the stage of concept popularization and experimentation. A lot of people started to know about cloud computing around 2009 and began to understand the concept of cloud. For the first 5 years, everyone was still understanding the concept and trying to combine cloud things with their own business. Of course, there are also some IT practitioners who are looking at how to find some opportunities in the cloud market, and the user's acceptance of the cloud is relatively low. The last 5 years are more about how to apply and implement. In the past 5 years, we have understood what cloud computing is, the key is how to use cloud computing, what kind of strategy to adopt, including the private cloud, hybrid cloud, public cloud mentioned earlier, which business is placed on their own platform, which is placed on the public cloud, and what is the hybrid cloud strategy, which has become a major topic in the next few years. But throughout this 10 years of development, I think it is really disruptive, and I obviously feel that traditional IT can be said to have been subverted by cloud computing. Cloud computing is equivalent to giving new players a chance to shuffle. Now the industry's largest technology companies, the biggest technology conferences, are basically talking about the cloud, which fully shows that the focus has changed a lot. That's why I started by talking about why cloud computing isn't a new bottle of old wine, because it's a paradigm shift, not just a certain technology point. A single technical point may not trigger such a big change, but more importantly, the entire computing model has changed very much, just like from cs to bs. In the 10 years from 2009 to 2019, cloud computing innovation has emerged in an endless stream, completely subverting the traditional IT industry, and I believe that the next trend of cloud development will continue to move forward and accelerate.
infoq: If you look at the progress made in the commercialization of cloud computing, what do you think have been the big changes in the past decade?
Fang Guowei: In 1943, ibm founder Thomas Watson felt that the demand for computers in this world was not so great, maybe 5 units was enough. I think in the last 10 years, you can see the development of the whole cloud, its concentration trend is obvious, maybe the world in the future is likely to be a few clouds to support most of the business, each cloud can be considered a computer. This trend has been around for the past 10 years, and maybe it will be the same in the future. On a larger scale, a data center is a computer, and several data centers are made into a cloud platform, just like the previous computer clusters. From this point of view, perhaps Thomas Watson's statement that there are only 5 computers in the world really has a certain truth, and I am more optimistic about this direction, and I am more convinced of the arrival of this day.
infoq: What will the cloud cloud look like in the future (in five years)? How much room is there for its development?
Fang Guowei: In 5 years, we can see several changes in technology: First, the hybrid cloud, multi-cloud model will be more popular, which is a very realistic approach, and this trend has been obvious in the past year. Users are very practical, and cloud vendors will be very practical when providing cloud services, just as we will provide different models such as public cloud, private cloud, private cloud, hybrid cloud, etc. for customers to choose. The second is that in the next few years, from the end, the edge to the cloud will be completely open. As the iot trend evolves, the pressure on the entire computing and storage will be very large. In the era of mobile Internet, a lot of information is triggered by people, but in the iot era, a lot of information and data are collected by various devices and sensors, they are 7× 24 hours of work, so the amount of data will change by orders of magnitude, which needs to be computed from the end to the edge to cloud computing, the whole chain is open, I think this will be slowly improved in the next 5 years. Whether it is a profession in cloud computing or an industry, it is still far from the ceiling. Why? Adding up the sales disclosed by various domestic cloud vendors last year, it will be found that this market is very small relative to the IT market. They may add tens of billions, and the IT market is at least hundreds of billions or even trillions. If we firmly believe that the cloud will gradually replace the traditional IT market, this space is obviously very large, and I think it should be difficult to hit the ceiling in the short term. And there are many opportunities in this market, and an important reason why Ping An wants to do the cloud is that this market is very large, the potential is also very large, and it is very important for Ping An's technology transformation strategy. I believe that other cloud vendors are the same, perhaps some of the Internet industry on the cloud proportion is already very high, but for the traditional industry, their proportion of cloud is very low, there are a large number of blue ocean markets need to be developed.