Shulou(Shulou.com)06/03 Report--

Is cloud convergence really coming? Listen to what CTO have to say.

Under the cloud fusion, how to break the technological innovation? Listen to CTO's motto.

The cloud has been shouting for many years, and for many ordinary technical people, there are a lot of questions in mind: what does cloud integration mean, what impact does cloud integration have on the company's technology system, and where will it go in the future? what technological innovation opportunities are happening and will happen? At the 2016AppCan Mobile developer Conference, five CTO and CEO, who are at the forefront of technology and user needs, had an in-depth conversation to talk about the past and present lives of cloud convergence.

Sharing nearly 9,000 words, I believe it will give you some inspiration:

Chen Qiwei, independent CIO and editor-in-chief of Mobile platform

Prologue: we are here to do mobile development, mobile ecology, as well as big data, cloud testing, intelligent voice technology, together is a small circle of mobile Internet, BAT is a big circle, we are a small circle.

Last year I wrote a book, Mobile platform, in which there was a paragraph about Mobile Internet 2.0 and talking about mobile cloud platforms + smart applications, but there was only one real core technology: cloud convergence.

Cloud convergence, by my own definition, is from the perspective of software. When users use mobile applications, the application will intelligently call the required networks, services and resources from the cloud according to the user's preferences, device conditions and scenarios, and provide them on demand. The back-end cloud management and application are naturally extended to the mobile end.

[question 1]

In fact, the cloud has been shouting for many years, and now the cloud fusion is really coming. Please talk briefly about the impact on your field and technology in the past two or three years. Where the future will go.

Zhao Qinghua, CTO of Zhengyi Mobile: integration of development and operation under cloud convergence

As far as mobile development is concerned, I think mobile apps are just talking about this iceberg. The rest of the 9x10 is underwater, and this part is what the cloud does. There is a successful woman behind every successful man, so there must be a successful cloud platform behind every successful application. For developers, if they want to make a perfect application and reach tens of millions and hundreds of millions of people, they must rely on the cloud platform.

AppCan believes that the functional scope of the cloud platform is very detailed and professional, and it is impossible for us to complete all the cloud functions. We only focus on the main line related to the development work, and we have also introduced other parts to other company products, such as Testin cloud testing, which we are also using to reduce 90% of our work through capability aggregation.

We focus on the development field. In the past two years, AppCan has been improving the cloud line from development to operation and management, providing integrated solutions to developers and enterprises. In fact, with the development of technology today, we need the support of system, platform and operation. What we are practicing is the integration of development and operation under the trend of cloud convergence, which is also the main line of AppCan.

Xu Kun, president of Testin cloud testing: each industry will be more and more subdivided.

We are a company that does cloud testing, what Testin does, what's more, you finish the development, and the rest of the quality assurance is provided by the company that specializes in testing. After working for five years, we found that developers began to use a variety of SDK and third-party services to reduce development costs, and developers slowly focused on the core products. We think that the whole society will be more and more professional, each industry will be more and more subdivided, entrepreneurs will focus more on their own areas of entrepreneurship, while other IM, statistics will have more professional companies to do it well.

Yun Zhi CEO Yin Jin: the first opportunity to challenge China's traditional enterprise software market

The seat arranged by the host today is particularly interesting. The front AppCan does the development, and then the Testin does the testing. This is related to the first two links of the App development. In fact, the challenge has just begun for everyone to develop, test and organize the APP online.

Because each App needs a set of business to provide support, the operation and maintenance of this business is what Cloud Wisdom does. You have a good App developed to ensure quality through testing, while operation and maintenance should continue to pay attention to user experience and business transformation. There are many indicators to see what you need to improve or why it will decline.

Cloud smart monitoring covers the mobile end and the back-end that supports the operation of App business, such as server, network, storage, and so on, all of which ensure the smooth operation of front-end App users. This process requires good monitoring and accurate data analysis. In the course of operation, this matter is very complicated and probably did not make any mistakes, but there is something wrong with the called third-party interface, which can also lead to problems with App.

After passing the hurdle of Cloud Wisdom, and TalkingData will do operational optimization and analysis, iFLYTEK will do better AI things. In fact, the five of us basically cover an App, from an idea to the last few steps of running it well.

Back to the host's question, Cloud and Duan, I feel very deeply, because Cloud Wisdom has been fighting in the field of application performance for six years. Six years ago, what we did was a website, and there was no App at that time. The Internet entrance of the enterprise is mainly the website, so we mainly monitor the performance of the website.

The mobile trend began in 2013, and the biggest challenge is the mobile's subversion of the traditional IT. The traditional IT is an internal matter of the enterprise, and the development, testing, and operation and maintenance are all internal systems, such as banks. A few years ago, we went to the bank to do business, either at the counter or ATM. To put it bluntly, it is a local area network that the enterprise can control, or at most a metropolitan area network. All problems are relatively easy to solve, because it is in a certain environment to solve a relatively certain problem.

But what do cloud computing and mobile Internet bring? It opens the boundaries of the entire application system, and you don't know when and under what circumstances your users will use your App to get in touch with the entire application system that supports your business. What this uncertainty brings is that you don't have a good way to manage the original problem that is manageable in a certain environment. So we need a new development model, this is what AppCan brings, we need new testing methods, we need a new data era operation and maintenance system, which is done by Cloud Wisdom, including the later operation and management.

I think this is what we are talking about today, the new mobile and cloud bring us technical challenges, but also new business opportunities for us startups. For the first time, we have the opportunity to challenge China's traditional enterprise software market, and we can do things in a new market. This is the business opportunity and value brought to us by the changes in cloud and end technology.

TalkingDataCTO Xiao Wenfeng: data inevitably requires strong convergence between the cloud and the end

TalkingData is a company focused on big data, providing developers with data solutions, including App statistical analysis platform, game operation analysis platform, advertising monitoring platform and so on. At the same time, we also do some traditional offline industry solutions, including banking, insurance, securities, real estate, retail and so on. When it comes to the integration of cloud and end, TalkingData actually feels the deepest. I also did App when I started my early business. I was an Android optimization master at first, that is, Baidu Mobile guard, and then I did Mobile guard and cleaning master in 360. I think for tool software, it can be divided into three stages: first, the function-driven stage, that is, the product manager crazily piled up functions on the software, whether good or bad, to see a competitive product with new features. Hasten to add this function, resulting in a software with more than a dozen large entrances and more than 50 small entrances, and customers can not tell which function is good and which is not. But in the early days of the mobile Internet, this was a competitive strategy, and for novice users, he saw more features rather than depth. At this time, App can basically run naked out of the cloud without having to interact with the cloud too much. However, with many functions, this also brings a problem, that is, the team's ability to focus is not strong, and there is no way to form a breakthrough at a single point. This was reflected in some super tools App at that time, and every function was done, but it was not a bright spot. Later, when it came to the 360 cleanup master, we wanted to focus on some of the functions to do more deeply, and data-driven was needed for cleaning acceleration, which started the data-driven phase. At that time, we found that cleanup acceleration was the most widely used feature, so how can this be done more deeply? We want to say that memory cleaning and disk SD card cleaning are provided to users, but novice users do not understand the concept of processes and directories. At this time, we wonder if we can directly provide users with some expert experience. It can help him identify which are redundant processes and which are directories that can be cleaned. At that time, we prepared a data operation team of more than 20 people to download a variety of applications from various application markets and install them on our mobile phones, resulting in a variety of directories and processes for analysis. It is very hard in the early stage, mainly artificial, and in the later stage is the way of artificial + intelligence, the efficiency has been improved a lot. When we put this experience on the client side, as soon as the customer opens the software, they will know which directories are redundant and can be deleted, and which processes do not affect usage and can be cleaned up. At this time, the interaction between App and cloud is already very close. Data will be collected on the end, desensitized and transmitted to the cloud. The cloud will analyze and accumulate experience through some algorithms or manual methods, and send the experience to the client at the same time. But does this mean that the depth of the product is enough? No. Because later we found that there were still a lot of customer feedback about why I cleaned up the navigation software when I was walking on the road, or why I cleaned up the health software when I was working out. Because at that time we only have data and experience, but do not understand the situation, do not know what the customer is doing, so it will lead to some misjudgment. Experience is static and it is difficult to avoid mistakes. Unfortunately, there was no means to help us judge the situation at that time. For example, the user is also walking, he is walking on the road, and walking in the park, gym, the scene is completely different. Also in the park, whether this person is an outsider, a local, an adult or a child, the scene may also be different, whether we can help users clean up more things according to different situations.

So this enters a scenario-driven phase. At this time, there will be stronger interaction between the client and the cloud, not only the interaction of data, but also the interaction of algorithms and intelligence. It will integrate more closely at this time. Therefore, I think integration is no longer a question of whether we want it or not, but that if we do not integrate, we will have a dead end. I think all applications in the future will be data-driven and scenario-driven, we will pay more attention to the operational level, operation can not be separated from data, data will inevitably require a strong convergence between cloud and end.

Wu Jiangzhao, Technical Director of iFLYTEK Open platform: there is actually no obvious boundary between Cloud and Duan.

At present, we have 150000 developers, and the number of App applications is about 120000, and the number of services provided to the outside world every day is of the order of 2 billion. Because we are the first in the world to provide voice delivery technology, so in China and even the world, is the largest voice cloud technology provider.

With regard to the cloud, my view is that there is no obvious boundary between the cloud and the end, the most important decision is the cloud or the end, there are two points, one is the technology, the other is the business requirements. Technology: if the technology advances and can meet different business needs, the business requirements can in turn promote the progress of the technology itself. In the past, things done on the cloud can be done on the end. Take iFLYTEK as an example, speech recognition used to do some in-depth neural network computing, which requires a lot of server resources to do on the cloud. But now, with the progress of technology and the increasingly powerful mobile phone hardware, iFLYTEK speech recognition can also be done well on the end, which can meet our daily life. In terms of home control, the limited scene command control is basically the same as that in the cloud. So the first point, in the past, it could be done on the top of the cloud.

In addition, what is on the end will also be put on the cloud. How do you understand this? Like the popular VR technology. There is a lot of data on VR that needs to be calculated and processed, which cannot be done on the end. However, due to network reasons, it is impossible to put such a large amount of data, even if the use of compression and other technologies can not put enough data on the cloud. However, with the advent of 5G or more technologies, it is no longer a problem for us to achieve data transmission, so the data on the end can be calculated in the cloud.

So from my point of view, there is no clear dividing line between cloud and end in the future. And now there are some new technologies, such as networking technology, maybe one day the mobile phone itself will be a small point in the whole big mobile Internet, and you will also be charged with helping others transmit some information at the same time, which is similar to the concept of networking, unlike now each mobile phone can only be individually connected to the telecom base station to get more communication.

[question 2]

Technology is changing so fast that Internet bosses are "treading on thin ice and trembling", and it will be more difficult for developers who are weak on their own. In your eyes, what are the similarities and differences in the demand for technology between developers and enterprise customers? According to the needs of developers, how to break through in technology.

Zhengyi Mobile CTO Zhao Qinghua: integrate enterprise-level experience and share it with developers

What does AppCan do for enterprises and developers? We hope that AppCan can filter before the developer chooses the technology. Because there are so many technologies, there are not many technologies that are really suitable for you.

AppCan is not only studying its own technological development, but also a special team is looking at various technologies of the Internet. we are studying many technologies. AppCan will deal with and try how to use these technologies and whether they need to be used in our platform. After you try it, the product launched by AppCan will tell you what technology you need to make this product application.

At the same time, we are also studying these new technologies, what goals we may achieve in the follow-up problems faced by the enterprise, we will uniformly package and organize according to the needs of the enterprise, and tell the enterprise how to use these new technologies.

Xu Kun, president of Testin cloud testing: developers are willing to try new things, enterprises seek stability

We all start with developer services. In 2015 and 2016, everyone will slowly move towards enterprise service. Because the biggest difference between enterprise service and developer service is that individual developers pay more attention to speed and efficiency, while enterprises pay more attention to benefit and quality.

To do corporate service, you should pay more attention to your own service quality, including communication and service with customers.

For developers, it is more online communication to provide more new things for developers to use and try. For enterprise services, more is the need for stable, high-quality service, or even 7 × 24 hours to ensure sustainable delivery.

Testin Cloud Test provides services to help developers find possible major defects. We have a large testing platform and a complete automation system to help developers test results on thousands of machines, find more testers to test your product, and the final result is that you may still have a bunch of problems, and enterprise personnel will fix all the problems, while individual developers will fix them. As long as the problem is not fatal, it can be online. So our feeling is that these are two systems. Individual developers are more likely to try new things, with high tolerance, while enterprise tolerance is very low, but you are willing to take good care of him, and he is willing to grow up with you and be a better partner in service together.

We are divided into two parts: a new product business will be more willing to be provided to developers, and when it is stable, it will become a paid enterprise service to sell enterprise customers, and we also have one-on-one account managers to communicate with you and ensure quality. When moving towards enterprise-level service, the challenge will be higher.

Yun Wisdom CEO Yin Jin: be more open and provide core competencies to developers

Individual developers and enterprise developers, many concerns are the same, for example, we are the pursuit of efficiency, the pursuit of quality. If fundamentally, why are we doing this? Everyone is for input and output. Personal development of some applications, hope to invest the least time and energy to achieve the goal; for the enterprise's goal is to invest the least capital, labor costs to achieve the goal.

In the era of mobile Internet, both individual and enterprise developers need to have better and ready-made tools or services to help achieve this goal. For example, after listening to Mr. Zhao, there is an impulse to write our company's own OA based on the AppCan platform. As a programmer, I am quite interested in this. This is a great input-output ratio, and a little bit of time and energy will produce a lot of output immediately.

For service developers, the first is to provide a lot of basic things, and the second is not only to provide a product, but also to open this capability to developers. In the cloud data era, for example, Cloud Intelligence has servers for active monitoring, which allows developers to call the global monitoring network with one line of code through API.

In the future, we need to be more open, provide core competencies to developers through API, developers can do what they want on the basis of basic capabilities, and achieve their goals with minimum cost and cost, which is what we enterprises hope to do.

TalkingData CTO Xiao Wenfeng: provide tools to accelerate iterations and adapt to new changes

Now the change is very fast, the environment is changing, the technology is changing, the data is changing, and the customers are also changing. How to deal with this change? I think today is still very sad, there are so many small partners, although the developer community is also quite difficult to make money, we all agree to earn money from industry customers to subsidize. There are still so many partners insisting that at least the original intention of serving developers has not changed, and we still hope to provide the best tools to help developers accelerate iterations and adapt to new changes.

There is a saying inside TalkingData that if we don't work as migrant workers, we should be contractors if we want to. What do you mean? Is to use more platforms, third-party tools to quickly achieve our iteration, to learn to "let others work for me," and focus on what you are better at. For example, if developers want to know their customers and who they are, they can use TalkingData's analysis tools and let TalkingData do the work for you.

In the future, TalkingData will focus on two things: one is to let the developer know his users and increase the customer dimension. For example, context awareness is our future direction. Last year, we did a context awareness marathon, which allows developers to know the status of mobile phone users, such as running or driving. Although this is only the simplest scenario, it enhances people's perception of mobile phone users. The other is to help speed up the iteration of operations. In the past, there was a problem with App operation, especially for operators. It took a long iterative period for me to do an analysis and bury a point. I asked the developer to bury this point and ask the product to assign you an event identification code. At the same time, I also asked the data department in the background to open the data interface of some products for you. You have to go through an online review of the market before you can finally go online and see the data. Last year, we released a new feature, "Smart Analysis," which basically freed the operator from this dilemma, allowing him to configure the location without code at any time, and see the data immediately-- which is equivalent to on-demand analysis. greatly improve operational efficiency. So TalkingData will make developers see more and see faster in the future.

Wu Jiangzhao, Technical Director of iFLYTEK Open platform: meet customer needs first, product promotion second

When we connect with to B customers, he pays more attention to what his enterprise qualifications are, what things and projects he has done, which projects coincide with his current needs in specific business scenarios, and he will also pay attention to the final development of these projects, and even ask us to send some demo, or take a look at the demo of our competitors.

On the other hand, developers pay more attention to whether your capabilities and features are novel, whether they can meet its needs, whether your interface is easy to use, and whether the document allows him to do the product as quickly as possible, and the stability will also be concerned about. However, unlike the enterprise, it will not form a contract that explicitly requires you to guarantee his service for 365 days, and there is no quantifiable standard that goes wrong. This is the difference between developers and enterprises that I see.

The second question is that when making specific products, we will generally understand the needs of customers. Just look at his application scenario, and then we recommend its technology, and we won't say that they need speech recognition to recommend him to use it immediately, which is irresponsible to the customer. Under the understanding of customer needs, work with customers to do a good job of App, and give a more complete solution.

As Apple said, in fact, the user doesn't know what he wants, he just knows that he wants it because he wants to solve a problem, but it doesn't have to be done this way. Therefore, in the area of customer care, we also put customer needs as the first place, not just to promote our technical capabilities.

[question 3]

Under the trend of cloud convergence, will the technologies and products of various companies be connected and integrated? Whether this kind of big connection and big opening can be regarded as a kind of technological innovation.

Zhao Qinghua, CTO of Zhengyi Mobile: openness is the premise of integration

When we are working as a platform, we should first ensure openness. Only by ensuring openness can we integrate others and let others integrate you. This openness, one is open to third-party platforms, the other is open to developers. Openness to third-party platforms is critical for AppCan, and we often face the final needs of our customers. When AppCan products are not fully satisfied, we will tell customers which products you need to purchase and do with AppCan. We are facing major customer manufacturers, he asked you to appear as a system dealer level, you have to provide other products. For example, when we tell you about Testin's products, customers read Testin's report. The end customer doesn't care about what you use, he only cares about cost accounting, and meets his requirements, and you are responsible for the consequences of the contract you sign.

We hope that developers can make better use of AppCan products, and at the same time learn AppCan technology, we integrate with each other.

Xu Kun, president of Testin cloud testing: gain revenue while helping users grow

We have been working with AppCan since 2014. We provide API interface to AppCan, and developers don't need to jump out of AppCan to get Testin test results. Testin has similar cooperation with many platforms, and we are willing to embed ourselves as a provider of capabilities into various development platforms to serve developers together.

I think we are all partners and our customers will share it. When he needs to do monitoring, I can push Cloud Wisdom, because I just do the test well, AppCan will do the development tools well, and we can recommend it to iFLYTEK when he needs SDK or voice in the future. We serve the same crowd, we will give better service to the user, then he will have a better perception of you. We are eager to share the user, for the user, he has another ability, for us, with the various partners, we have another user. So basically we will do this kind of integration. What we hope more is that when you have deeper needs, I think developers are really different from enterprise customers. Developers become enterprise customers only when they make money. When you need better service, whether you can buy my enterprise service or not is the common survival logic of our enterprise. So in helping users grow, everyone is willing to share. The whole market plate is getting bigger. We do corporate services, we can get our income.

Host: the future of the mobile market is infinite. It can be done in increments.

Xu Kun: we hope that developers can focus on its field, serve their users well, make their experience better, their business model better, and make money. If he doesn't make money, why is the tolerance so high? it's because he has no money.

We hope to create an ecological environment, we work together to build an environment, so that developers slowly become corporate customers, after becoming corporate customers, his consumption concept will change. We want him to know me first when he does the development. This has always been an idea of Testin. So we will work with all platforms.

Yun Wisdom CEO Yin Jin: individuals achieve the extreme and integrate with each other

I think when faced with this problem, what we need to do is to do open integration with an open mind, and focus on doing our own things well and providing better services when the times are changing. We all have the unique ability, hope to achieve the extreme at the same time can open this ability out, no matter be integrated or integrate others, we are doing this, so that we can build a better platform and ecology for service developers.

TalkingDataCTO Xiao Wenfeng: the data itself is very difficult to establish exclusivity, open to survive

Developers must look to the future with a more open mind, especially for companies that make data. At the beginning of this year, we attended the Strata Hadoop Technical Conference in the United States. We think that big data's technology is a little surplus to some extent, the same scenario can be solved with a lot of technologies, and the data itself is very difficult to establish exclusivity, so in order to survive in this case, big data must remain neutral and face the market and customers with an open mind. We all have cooperation among these companies today, and they are all entrepreneurs. We know very well that it is very important to have an open mind, so this is the core point that we share together.

Wu Jiangzhao, Technical Director of iFLYTEK Open platform: open collaboration to expose developers to the latest technology

As a capability provider, we need to work together to make it easier for developers to start a business. To take a simple example, identity authentication, there are many means, face, fingerprints, body prints, iris and so on. IFLYTEK mainly does cloud technology, we can provide voice prints. We cooperate with Shangtang Technology, they provide faces, so the combination is a good way to solve the identity authentication problem, which can be solved for most mobile phones. So in terms of capability, capacity providers should work together to give developers access to the latest technologies.

Second, talking to TalkingData about data, because now each enterprise is limited by the services and independence he provides, the data collected can not be comprehensive, future developers, manufacturers, and ultimately serve real users. If you want to serve him well, you should know all kinds of information about him, including his hobbies, educational experience and so on. At this time, we need all our companies with data to share each other's data and build a large data warehouse in order to get the real information of this person and make this person's user experience better. Of course, this is a relatively long-term thing.

Compere: today is a small mobile Internet circle, we talked about the integration of technology, cloud and end, that this is a trend. In addition, we all believe that both capacity providers contribute our capabilities, share data, integrate the industrial chain, make our developers more efficient, make money as soon as possible, and solve the problems of the enterprise. this is the ultimate goal that we have in mind.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.