At the sixth session of the cloud Computing Conference, Zhang put forward a new idea

The following is the vice chairman of the China Electronic Society cloud Computing Experts committee, vice chairman of China Large data Experts committee, Microsoft Global Senior vice President, Microsoft Asia Pacific Research and Development Group Chairman Zhang Transcript:

Zhang: Dear Minister Wu, you leaders, ladies and gentlemen, good morning! I am delighted to be at the annual cloud Computing Conference, this is the sixth session, I am honored to participate every time, and have the opportunity to report some progress in cloud computing, especially the Microsoft Company's progress. In fact, it is also an opportunity to think about how the cloud has developed, what problems it has, and what opportunities it has.

I remember the cloud computing conference in the previous years more still say what cloud, cloud is a real thing, or a kind of flicker? The definition of cloud is always in an educational state. Later went to the application, in the end how to deploy the cloud. Now has become the industry consensus, all it enterprises, the global use of the cloud, whether it is as a cloud platform, services, applications, are using the cloud. Google, Microsoft, Amazon, IBM are in the large-scale deployment of cloud, China basically all it internet companies have a cloud, I often joke with friends, now which enterprise without cloud, are not very good meaning, it can be said that the cloud has become a real trend of technological development, but also an important deployment, The cloud is on the path to benign development.

I would also like to take this opportunity to report on the development of Microsoft in the cloud, including our global deployment, especially in China. In addition, I would like to take this opportunity to talk about some of the major trends in Internet development, one of the most important engine and driving force is the development of the cloud. What I want to talk about today is the physics of cloud computing leading the internet, and I'll explain what that means in a moment.

If you look at the development of the Internet, I think it is very interesting, this year is the internet in China 20, in the world may be about 35-36 years. In the 30, what we did was to virtualize and digitize a simulated world, or a physical world. Your text, your music, video, business processes, information to digitize it, and then use the Internet to connect it together, people and people connected to a social network, people and objects, goods linked together is the electric quotient, people and information is connected to our various browsers---Internet Information platform. So, in the beginning, it was primarily a time of physical to virtual virtualization.

Now I think we are moving in another direction, that is, the Internet is actually entering the real world from the present virtualization and moving towards physical. The concept of physical is not a very new thing, many enterprises have started this aspect of deployment, research and development for a long time, this two years, especially the past year has made great progress.

What is called the physical of the Internet? I summed up the situation in this industry. There are actually three aspects of this, the 1th of course is to expand the technology of the Internet to the real world, including our sensor network, smart home, Smart cars, smart grid, wearable computing, is the use of sensor networks, cloud technology, using the Internet technology, are from this level and the circumference of understanding. For example, many enterprises in the beginning of this deployment, such as IBM used to speak of the wisdom of the Earth, GE talked about the industrial Internet, the Internet technology into the transport, power grids and so on, into the real economy.

Another circumference, it is important that the Internet business model and thinking way, now a very popular word is called Internet thinking, I think it is very reasonable, this way of thinking whether for Internet enterprises, it companies or traditional enterprises, there is a great revelation, which includes the user first, rapid iteration of your products, Fan economy, viral marketing, in fact, the most important thing is to have more direct interaction with users, and quickly to make the market response. In here is indeed and our traditional enterprises have a big way of thinking different, to the user first, with users, with traffic, users like things, to keep it, provide more value, and then to make money.

The other one, as I mentioned earlier, it is true that the Internet is entering the traditional industry, and sometimes it will impact and subvert the traditional industries, from the past media to retail, now into the financial industry, into the education industry, into health and other industries, will be the Internet technology, thinking, business model changes, shocks and subversion.

I'm not going to tell you this, I've been writing an article recently, when it comes to the impact of these three dimensions, I think that the trend for China, whether global or domestic, will give us a lot of opportunities, and in some ways China may lead the United States or other countries, Because China has a large traditional physical manufacturing industry, we have a large market size, user size, talk about big data, talk about the whole scale effect, talk about how to upgrade the traditional industries, China's demand is the largest.

And Chinese companies in many ways, this kind of open thinking, this entrepreneurial spirit may even surpass the United States. For example, electric business, in the United States, you often buy something for several days before you can, a taxi has a taxi software, may pay more money. And then in a lot of enterprise use things, because it currently has a more mature it architecture, so the development speed will be slower. So, I think the physical process of the Internet provides a great opportunity for China.

Just now I talked about the fact that there are many phenomena behind the clouds, we see physical, see a lot of sensors, see people and people connected, people and things connected, things and things connected, in the future the total number of intelligent devices will increase about 10 times times, this is a conservative estimate, there are 500 smart devices, The resulting data, whether in terms of volume or a variety of properties, will be explosive, a lot of real-time information, a lot of unstructured information, and so on.

This morning did not know that there were no academicians and experts to talk about, recently in some new developments in artificial intelligence and machine learning new developments, Microsoft has been in this area for almost decade years, 20, spent a lot of energy on artificial intelligence, especially machine learning, human brain research, now there are a lot of breakthroughs, Especially deep learning. As you may have seen recently, Facebook, Google or Microsoft are all doing a lot of research on this, and Baidu also hired a top scientist in this area last week, and he will lead Baidu to do research on this.

This is actually for the future of the Internet large data processing, for the machine more intelligent, artificial intelligence development has very important significance. In the past we have become a very cold thing for many years, making things because the rules of artificial intelligence are of little use. Over the years, with more and more data, and with more computational power, including brain recognition, we have made a real breakthrough in artificial intelligence in this area. I often tell the students now that machine learning is as important as your English and your computer operating system, is a required course, and for your future career development will have great benefits.

A brief talk about what we've done recently in Microsoft Studies is to use deep learning to help them make a speech smart. In addition, let it to cognitive pictures, we know that cognitive image is a difficult thing, early through a variety of so-called characteristics, to find the characteristics of the image, and then to do a match, do a search, do some learning. Now you can use a multi-level deep neural network, we do a 9-layer, we do almost thousands of GPU set up. You find it very interesting, at every level of learning, the machine is looking at the beginning of the same shape as people, to the back you find each layer out, it is more and more like the brain, and finally find or face, the real object is found out. So, the development of this aspect really can be as smart as people. At present, it is possible that almost 5-6-year-old children have such brain intelligence level, in 10 years, 20 years or so, will reach an adult's intelligence level, and the future will certainly surpass mankind.

The question now is what do we do if the machine's intelligence really transcends human beings? Last week we were having this conversation about these issues, the impact of deep learning on the future of artificial intelligence, which is a huge boost for cloud computing, large data and physical physics, otherwise, without such a big development, we have so much data is useless, It's no use concentrating so much data, just doing some simple digging and analysis, the real data is going to make your machine smarter through the cloud, learning how people interact, learning how people think, how to judge, and being a real helper, and one day the assistant is smarter than you.

Now our data center has expanded to 16 locations in 8 regions, with a lot of data centers, and we've invested $15 billion in infrastructure. Just now, regardless of storage or calculation, and the rapid development of 55% of the Fortune 500 are using Microsoft Azure, we double the amount of calculation every 6 months, every 3 months storage is doubled.

We're different from other cloud computing services. 1th, we can support a mixed cloud, a seamless link between a private cloud and a public cloud, no matter where you are, including your operating system, including your virtual machine management, system authentication, security management, development language, environment, Architecture, can be completely seamless links.

Of course, we think the development direction of cloud computing must be to the public cloud development, but in a long time, perhaps there will always be a mixed way, I think this is a big advantage of Microsoft, we are the only enterprise can provide seamless, large-scale mixed cloud service solution.

2nd, we provide three levels of service to the enterprise, IaaS, PAAs, and SaaS.

3rd is also a lot of customers do not understand, we are fully open to support open source solutions, including the operating system, we support a variety of open source development tools, whether it is PSP and so on, we will support. In addition, we also support open source large data platforms, including Hadoop.

In China, we announced a public preview last year at the cloud Computing Conference, a year after the announcement of formal business, a year of rapid development, in fact, many aspects have exceeded our expectations. The way our Microsoft Azure landed in China was to connect with the century interconnection, to be operated by the century Interconnection, to provide technology by Microsoft, so this is a good model, with the world's most advanced cloud computing architecture technology, plus our local size, services and operations.

Another is Microsoft Azure in China's code name is "Moon Cakes", this is three years ago the mid-Autumn Festival when we design an idea. It supports all the features that I've just talked about Azure in the world, in addition to having two data centers in China and providing a backup of 3x2 in Beijing and Shanghai, which provides a wide range of services, I remember we opened 5,000 accounts in the preview and ran out in 24 hours, Recently we opened a number of accounts, developed 2000, in 4, 5 hours immediately ran out, so we are now basically in the stage of demand, some companies are multinational companies want to provide cloud applications in China and its services, Some of our domestic businesses want to use our domestic cloud including Azure in the global layout, and most of the other businesses are for domestic service. Therefore, security, privacy is fully guaranteed, and fully meet our Chinese government regulations and policies.

No specifics, I think a very important advantage is that our cooperation with the century interconnection, as a global enterprise, despite good technology, but often in the domestic provision of services, there are many standards, many operations, many maintenance, including rapid response is relatively scarce, Therefore, and the domestic leading like the 21st century interconnection, including the blue flood such business cooperation has brought us a great advantage, to achieve a very good complementary.

I'm talking a few examples of azure, one example being the 2014 Sochi Winter Olympics, which is a very complex and also a large-scale streaming media application, possibly the largest streaming media application ever. In almost 24 countries, more than 200 channels, hundreds of millions of people, which includes all of our coding, transcoding, playback, content management, streaming media, and a variety of interactions, is a very complex application, before and after three weeks. So, this also shows that cloud computing for real-time media can do large-scale applications, and it is the first time to verify that we also learn a lot of things, as a whole this is a landmark case.

The second, and this very similar, is at home, we use the domestic azure to support the Spring Festival Gala, although it is at home, but this size and the Winter Olympics are also similar, time will be shorter, within 24 hours there are hundreds of millions of people to watch, but also to interact. And this is a sudden event, less than 24 hours later. So we cooperated with CNTV to ensure the playback and interaction of the Spring Festival Gala on the network.

This is a good example of a mixed cloud, in fact, CNTV content is in its network, our azure provides analysis, calculation, playback is in our cloud above, both sides of the interaction and coordination is also very smooth. In addition, CNTV is open source, so our support and compatibility of open source is relatively successful. Cntv's CIO told us that using Azure, the entire deployment time from three months to three days, reduced the cost is 90%, the efficiency has been greatly improved.

Another case is that Office365 is a project with the Shanghai municipal government, which is our service on the cloud and the largest SaaS application on Microsoft Azure, the fastest-growing service in our entire business, the first to reach level 1 billion revenue. At home we are glad that we have been used in the Shanghai municipal government.

I put the video below is another cloud, is to do the cloud game, we also know, now at home is also a big trend. It's also the first time we've been on azure to support a massively high fidelity, highly interactive game Titanfall, the biggest cloud advantage ever, fully implemented on Azure.

Not only are big companies and small companies using the cloud, but many startups are using the cloud. We've done one thing in the last two years, my own excitement was that we worked with startups to do an incubator-like accelerator mechanism, which has nurtured more than 60 start-ups in two years, and these companies are using azure,azure to improve their overall business development. We invite these companies in, and then provide them with cloud, technology, entrepreneurship counseling, including funding links to help them to their products and services to the domestic users, to the global users.

There are several figures, these companies are now more than 100 million users, the market value reached 1.5 billion U.S. dollars, 6 months, the market value increased by about 5 times times. In addition, I am also happy that these companies provide us with a lot of feedback, especially when the cloud landed in China, there are many performance aspects, technical aspects, they are helping us to explore and solve.

Let me briefly talk about the Internet physics, what Microsoft is doing. Some people may recall that Microsoft made a wearable calculation and a smart watch early on, which we did more than 10 years ago. Then I made a smart fridge, smart cups, so we have been very optimistic about the IoT, but at that time too early, feel that the time is now more mature, so we have recently set up a department called Microsoft Things Networking Department, recruit a group of the best system design sensor machine learning experts.

Recently introduced is the Microsoft Intelligent System, which has how to collect a variety of real-time information, the information is very good collection and release, and then to do intelligent processing, using machine learning algorithm better presented to the user, finally feedback to some sensors above. There may be some based on rules, some based on learning and statistics, and finally connected to the brain, the brain is the cloud of Microsoft.

One application example is our intelligent City Citynext, this is our intelligent system large data Application best example, in fact, the city is a big data, we citynext platform the whole way of thinking, solutions and the past are completely different, we are a multi-level, and is a middleware, The middleware has more open applications, and is centered on people and urban actors. Last year I talked roughly, we were a project last year, and this year we have started deploying and applying in cities. I simply put a video to introduce our outlook.

We all know that Microsoft has changed a lot lately and we have new strategies, new deployments, and new CEOs. The new strategy is actually a continuation of our past mobile priority, cloud priority strategy, I just talked about all the concepts behind a big cloud, the front are intelligent devices, the key technology is machine learning, is a new generation of user interface, in fact, to achieve any device, anywhere, any application and any data, We call it ambient FDI, so people are connected to people, people and things, things and things.