'Cloud nativity is the only way for the development of cloud computing. It can well solve many pain points of cloud computing in industry applications. In the long run, the development of public cloud will develop towards a more open trend in the future, and it is unlikely to operate separately, otherwise it is not conducive to end users.' Recently, at the 2019 trusted cloud conference, Chen Yili, deputy director of the cloud computing Department of the Institute of cloud computing and big data of the China Academy of information and communications, said in an interview with the global network technology reporter.

It is generally believed in the industry that cloud computing has entered the pan cloud era, and cloud computing has also completed its initial mission, from the transformation of IT infrastructure to agile cloud infrastructure, into a new life cycle, that is, to establish a digital value chain and release the original vitality of business.

'As the basic base of current information system, cloud can integrate more service capabilities, from the underlying computing resources to the platform to the complete intelligent system, and can obtain business value flexibly, quickly and cheaply according to the actual needs of users.' Chen Yili pointed out that there will be some new development opportunities for Cloud Computing in the vertical field, and the core is to build a one-stop platform of AI technology with the help of cloud computing.

In his view, according to the development history of ABC (artificial intelligence, big data and Cloud Computing), the development trend of artificial intelligence and cloud computing is similar. 'Why is AI so popular until recent years? Because we lack computing power and data. The development of big data is relatively short, but it provides AI with algorithm and data analysis capabilities. At present, big data and cloud are constantly precipitating users' businesses and data. This time is' waiting for the cloud '. Cloud computing is gradually developing in depth, and intelligent cloud system is gradually taking shape.'

The pain points of AI on the ground commonly reflected in the market are: first, it can quickly learn and verify; Second, it can iterate efficiently; Third, the computing resource itself has a high threshold; Fourth, find the right landing scene.

Chen Yili believes that cloud intelligence may be a good way to implement AI. 'Cloud computing has significant advantages. First, it can quickly schedule resources according to demand; second, it has a sound service market, which can rely on third parties to build an overall ecosystem; third, it can expand and accommodate some resources according to needs; finally, the enabling platform in the middle PAAS layer can better enable the upper business. Therefore, cloud computing will usher in a new era of intelligence.'

Later, he explained in detail the two statements in the industry, 'cloud intelligence' or 'smart cloud':

'Cloud' is a technology path.

At present, cloud has become a new form of IT infrastructure. As an underlying facility, cloud has become a must choice as enterprises gradually deepen their cloud deployment. In the long run, the advantages of cloud are obvious. Although some core applications still rely on traditional physical infrastructure, the elastic expansion, flexible architecture, rapid delivery and agile product iteration of cloud cannot be achieved by traditional it. Cloud has also gradually penetrated into core and key businesses from the initial marginal businesses in enterprise applications.

'Intelligence', that is, intelligence, changes our mode of production.

Cloud computing is an important window. The three elements of artificial intelligence include computing power, algorithm and data. What can cloud provide? First, it can provide sufficient computing resources and massive data. Intelligent production mode can completely change the operation mode of the whole society and move towards intellectualization and automation. At the same time, the cloud releases huge productivity, improves work efficiency and accuracy, helps us make decisions, and realizes the digital transformation of the entire industry.

'Energy' is empowerment.

With the continuous improvement of the intelligent cloud system, it continues to export its capabilities and empower the entire industry. With the improvement of the ecosystem, it is constantly doing subtraction (de Wheeling) to help more small and medium-sized enterprises lower the threshold, and then focus on business applications more quickly. Because the development trend of cloud computing in the future is also towards mutual synergy, which can promote the innovation and development of the whole industry.

In Chen Yili's view, cloud intelligence also has three major attractions for the industry - cost savings, lowering the threshold and improving innovation capabilities.

He further explained that, 'First, save costs. A GPU host costs about 10000 yuan a month, and the cost is relatively high. The cost of buying a card is even as high as 100000 yuan. The form of renting instead of buying can reduce the threshold, especially for start-ups. It can make it easy for him to obtain such resources, and use them out of the box. Second, reduce the threshold. On the basis of basic resources, we can quickly build intelligent services based on cloud services Specific environments, such as building a model algorithm and reasoning process, can be quickly implemented on cloud services. Third, innovation ability. Face recognition, speech recognition and natural language processing based on public cloud services are relatively mature, which can be directly integrated into various business systems with the help of rich third-party services to stimulate industrial innovation. '

In addition, from infrastructure to enabling platforms to application services, Chen Yili, including the cloud computing and big data Institute of the China Academy of information and communications, hopes to gradually improve the Intelligent Cloud standard system. At present, the whole standard system is divided into three parts: first, infrastructure, namely CPU, GPU, FPGA and ASCI. The second is the enabling platform, which is oriented to data, data processing, reasoning process, machine learning and other capabilities. Third, application services, based on the intellectualization of image and voice, provide a variety of SaaS applications, and finally face the general public, making our life better.

He said, 'in the future, the so-called pan cloud or pan connection in the field of smart cloud, we may explore edge computing, the Internet of things, smart cities, and integrate with more industries.'

At the end of the interview, Chen Yili told reporters his outlook for the future of smart cloud: cloud nativity is a road that must be taken. Because industries need a mutually open environment, if they operate separately, it is not conducive to end users.

'First of all, we should open source and open. Software must be the cornerstone of the future information system, especially the cloud native, which has developed quite well in the past two years. It can solve the user's better landing and can realize the very active technical suggestions of digital transformation. Second, establish a synchronization mechanism between standards and technological development, so standards represent the problem of industrial consensus to a certain extent. Third, ecology and collaboration, for any technology or For industry, ecology is very important. At present, there are still high technical and talent barriers in the smart cloud, and the penetration rate is far from enough in the application scenario. We need to cultivate a good ecological environment, cooperate with various parties, and jointly promote development. ' He sighed.

Note: the concept of cloud native was first proposed by mattstine from pivot in 2013 and has been used ever since. That is, by adopting cloud native technology and management methods, businesses can be better born in the 'cloud' or migrated to the cloud platform, so as to enjoy the efficient and sustainable service capabilities of the 'cloud'. A consensus in the industry is that it is not a product, but a set of technical system and methodology, and digital transformation is the overall change of thinking first, from inside to outside. More specifically, it is a culture, a trend, and an inevitable direction of cloud computing.