With the diversification of application scenarios, "computing power" has also ushered in the era of "intelligent services"

With the development of the IT industry, "computing power" has played an important role in the "civilian" and "commercial" fields, which has posed challenges to the usability and compatibility of computing power platforms. With the diversification of application scenarios and the high integration of facility management software, "computing power" has also ushered in the era of "intelligent services".

Computing power is everywhere

As a data expert at an Internet company, Song Ge (pseudonym) has been working overtime a lot recently. His job is to enable the "artificial intelligence" system developed by the company to more quickly identify the "massive junk photos" uploaded by users.

With the introduction of the "strictest" garbage classification measure in history, the "Shanghai Municipal Domestic Waste Management Regulations" in July, Song Ge and his "artificial intelligence" joined the war between people and "garbage". Each of Song Ge's big data recognition models brings people closer to victory in this war.

On the third day after Typhoon Lekima made landfall in Zhejiang, Wang Rui (pseudonym) at the Zhoushan Meteorological Station in Zhejiang was busy comparing the ozone data of Cyclone Lekima collected by her colleagues from weather balloons. These data will be combined with temperature, air pressure, wind speed, wind direction, satellite, observation and other data to form one of the data sources for the prediction and analysis system for typhoon intensity, path, environmental damage, etc.

Satellite photos of typhoons

Zhang Zhiqiang (pseudonym) at the "Genome Research Institute" is comparing a marker gene of bacteria in the human intestine with more than 100 other types of bacteria to determine the microbial community environment of the fungus and help people understand the composition of bacterial communities in the intestine and related effects or hazards.

In six months, the new model will roll off the production line. Li Yuanhang (pseudonym) is adjusting the appearance of the new model to reduce wind resistance. He is conducting a "simulated wind tunnel" experiment on the appearance of the new car on his computer. If the experiment is successful, he and his team will make a 1:1 simulation model and test it in a real wind tunnel.

The work of Song Ge, Wang Rui, Zhang Zhiqiang, and Li Yuanhang all requires the support of a resource called "computing power." As the complexity of application scenarios increases, people's demand for "computing power" is increasing day by day.

Rapidly growing demand drives new trends

Taking the field of artificial intelligence as an example, in May 2018, the US non-profit organization OpenAI released an analysis showing that the demand for computing power has increased by more than 300,000 times since 2012. Among them, the computing power used in artificial intelligence training tasks currently doubles every 3.5 months.

However, with the demand for energy saving and cost reduction and the diversification of application scenarios, the deployment and management of "computing platforms" are becoming complicated. Compared with the deployment costs of tens of thousands of yuan and the construction and debugging costs of more than half a year, more and more companies and scientific research institutions choose more flexible methods such as leasing computing resources.

Advanced computing services that are "simple", "customized", "integrated" and "management as a service" are becoming the mainstream market demand.

Four major trends: "Computing power" enters the era of intelligent services

Trend 1: Simplicity

As the needs of industry users upgrade, computing power platform management software is now much more integrated than before. Not only does it need to allocate computing resources, but it also needs to monitor hardware conditions 24/7 to ensure smooth operation. With the upgrade of computing power infrastructure software, obtaining "computing power" will become more "intelligent" and "simple".

▶▷Trend 2: Fusion

In addition, there is a general demand for "artificial intelligence", "cloud computing" and "data center" in the industry. This requires further "integration" of "data center", "cloud computing" and "artificial intelligence" at the software level to meet the various needs of users.

Trend 3: Customization

In order to use computing resources more flexibly, enterprises will lease computing power more frequently. However, different fields have different requirements for the deployment of computing power, which requires customized configuration and in-depth optimization of corresponding computing resources for enterprises based on different industries and fields combined with different application directions.

Trend 4: Management as a Service

For a long time, providing "solutions" based on industry needs has been the traditional business model of the computing industry. However, as the demand for "simplicity", "integration" and "customization" increases, the core services of "solutions" will focus on the intelligent management of "computing power" from infrastructure to software environment. Managing "computing power" and making the use of "computing power" more "intelligent" is the ultimate direction of the solution.