Shulou(Shulou.com)06/02 Report--

Nowadays, the popularity of artificial intelligence has accelerated the throbbing of the cloud computing market, and the intelligent technology represented by AI has been incorporated into the territory of cloud services. Cloud computing from its birth to now, more than ten years of rapid development, is under the injection of AI ushered in a great change.

In the past, the construction of data center computer room needs to follow many norms and standards, including data center computer room classification and performance requirements, computer room location selection and equipment layout, environmental requirements, building and structure, air conditioning, electrical technology, electromagnetic shielding, computer room wiring, computer room monitoring and security prevention, water supply and drainage, fire protection technical requirements, etc., with long construction cycle, large investment, high operating cost and low efficiency.

Modular data center is a new generation of data center deployment form based on cloud computing and intelligence. in order to cope with the new development trends such as machine learning, high performance computing, distribution and elastic expansion, it adopts modular design concept. it minimizes the coupling of infrastructure to the computer room environment, and integrates power supply and distribution, refrigeration, cabinets, airflow containment, integrated wiring, dynamic ring monitoring and other subsystems. Improve the overall operational efficiency of the data center, achieving rapid deployment, resilient expansion, and green energy saving.

Modular data center can meet the urgent needs of IT business units for future data center infrastructure construction, such as standardized design, component factory prefabrication, rapid online deployment, effective reduction of initial investment, in-module energy pool management, high utilization of dynamic IT infrastructure resources, intelligent operation and maintenance management, and ensuring important business continuity. Provide shared IT services (such as cross-business infrastructure, information, application sharing, etc.), rapid response to changes in business needs, flexible capacity expansion and green energy-efficient data centers, etc.

At present, with the continuous popularity of artificial intelligence, its demand for computing power has long exceeded the development of Moore's Law of CPU. We can see that heterogeneous computing represented by GPU, FPGA and ASIC in the basic resource layer has become the direction and trend. The platform layer provides intelligent tool components and model training management, and there are more and more intelligent cloud application services on it. Turn heterogeneous resources into a universal AI computing capability through the cloud to achieve flexible supply, that is, business peak calls, business trough goes away, a large number of business challenges can be easily met, convenient intelligent cloud services can quickly meet the needs of production scenarios, and then promote industrial upgrading.

We know that according to the delivery method, public cloud is mainly subdivided into IaaS, PaaS and SaaS. Globally, SaaS is the mainstream, that is to say, the main capability of the cloud is to host different software and application services. In the domestic market, IaaS still accounts for more than 60%. In other words, infrastructure is still the main role played by domestic cloud services. This shows that the core goal of most Chinese enterprises that use the cloud is not to obtain an application, but to migrate their original business and applications to the cloud and expand the utilization of resources to replace the high cost of the server. In this case, the public cloud plays a more business-supporting role. With the integration of AI technology into the cloud landscape, it has changed the delivery of cloud services and achieved significant progress in optimizing productivity.

The core competence of AI is that it can replace part of the manual work with high repeatability and immediacy. For example, computer room inspection, security, quality inspection, human flesh operation and maintenance, there may be intelligent patrol inspection, intelligent security, intelligent operation and maintenance, unattended, self-driving and so on. It is not difficult to find that they do not need the characteristics of the cloud to support the business, but they greatly need AI to optimize productivity and achieve the goal of "improving quality and increasing efficiency".

The combination of AI and cloud services has produced many high-value services, and then a series of changes may take place from market behavior, business logic to service demand. This is the core reason why today's cloud service market is booming again, and a large amount of capital, leading enterprises and developers have begun to gather to do ecology.

Of course, today, the combination of cloud and AI is still in its infancy, and more still stays in the propaganda of manufacturers. To really sink the AI technology into the industry, we need to go deep into the production scene, combine the business to complete the scientific training of high-precision models and some complex operations, and have certain reasoning power on the edge side and terminal side, so as to achieve the optimal calculation-effectiveness ratio. This requires the AI computing power of the cloud plus side to be closely combined with the computing power on the hardware to build a whole. Therefore, the maturity of cloud + AI services still needs time and further development of technology. From a historical point of view, the end-to-network in the server era, and then the cloud-to-network in the cloud computing era, will become a new cloud-network-to-end structure in the AI era, which makes cloud manufacturers have to innovate themselves, and also allows the industrial service market to find new opportunities.

In the digital age, the combination of AI and cloud, so that the data is no longer the cold numbers placed there for people to watch, people have made it benefit and value through the analysis of the data, and changed the traditional enterprise cloud service model. We believe that the future of AI artificial intelligence is bright, but the road ahead is tortuous. As long as you go in the right direction, you don't have to be afraid of a long way to go. If you keep going, there will always be a process from quantitative change to qualitative change.

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.