Husaku is all hot, and millions of enterprises set up a new idea of "data house"

Source | Tech Planet

Wen | Jia Ningyu

If it is currently on the track of the investment field, "data intelligence" will definitely be considered fragrant.

In August 2021, the super unicorn Databricks received $ 1.6 billion in round H round of financing, with a valuation of up to 38 billion US dollars, a hundred times the valuation of 4 years. Another company Snowflake, the first day of listing in 2020, a market value of over 70 billion US dollars, becoming the largest software IPO so far, creating a return of investment in US stocks.

Looking back at the domestic market, data intelligence is also becoming a capital of capital. For example, data intelligent service providers Didpu Technology had financing 4 times a year, and recently received 110 million B+rounds of financing. It has grown into a new generation of lake and warehouse all -in -one data intelligent basic software field in only four years. Essence

Such companies are mainly studying the core products of data intelligent tracks, that is, analytical databases based on lake warehouses as the architecture. Data show that the market size of China's analytical database in 2021 was 24.99 billion yuan. It is estimated that in 2024, the size of China's analytical database market will reach 52.14 billion yuan.

"How to make the data -driven business and the value of the data truly release, I think it should fundamentally solve the problem of underlying basic software systems, and then consider the application of upper -level data." Planet, when the world enters the era of data bangs, the real internal force of corporate competition is the ultimate value brought by big data.

The market heated up again, why is the "integration of lake warehouse" fire?

Looking back at the earliest data storage processing solutions used by domestic and foreign companies, basically "data warehouses" built by Oracle and IBM for enterprises. deal with. Even some large enterprises invite IBM to do overall business consulting, and with a set of localized software systems to establish a standardized data warehouse. In 2002, Hadoop technology was also born to meet the needs of the company's basic data storage analysis.

However, with the development of enterprise business, data infrastructure of any form (including structured and non -structured) and any format (including text, audio, video, and images), as well as higher requirements for real -time proposal. As a result, DataBricks launched Delta Lake in 2016. This group no longer builds a "data warehouse" for enterprises. It follows the "data lake" route. Although it meets the needs of corporate structured and non -structured data storage, they do not do not. Support transaction processing, does not guarantee data quality, and lacks consistency/isolation. Snowflake's rapid popularity, its cloud -based data storage and analysis services provides a new way and ideas.

Therefore, there is a voice in the industry: Is there a technical solution to solve the disadvantages of data warehouses and data lakes and the advantages of re -integration? In 2020, the industry developed a node, and finally a better solution appeared.

This year, DateBricks first mentioned Lakehouse, which is the concept of "Lake Cang". "Lake Cang integration" is not a simple technical integration of lake+warehouses, but a new type of architecture that integrates data lakes and data warehouses and is more open. Some people make it a metaphor, similar to setting up many small houses by the lake, some are responsible for data analysis, some operate machine learning, some to retrieve sound videos, etc. As for those data sources, they can from the data lake. Get it easily.

Because of the advantages of the data lake and the position, and the new architecture for the AI ​​era, Gartner announced in the 2021 data management report that the "Lake Cang integration" entered the maturity model for the first time. At this point, the "Lake Cang integration" has caused greater attention.

In Yang Lei's view, the deep -seated reasons of "Lake Cang" are mainly because the volume of the data is large enough, and the demand for large -scale data to be processed in real time is even more valued. Another thing that promotes the development of Hucang is the large -scale popularization of AI machine learning. All walks of life urgently need to use machine learning algorithm to support the management and innovation of data.

Everbright compete, innovative competitions in the cloud era

In other words, the formation of the integrated structure of the lake is a market supply and demand. However, companies such as DataBricks have not expanded the domestic market, and tens of millions of domestic companies are facing the challenge of digital upgrades. Domestic data intelligent services manufacturers are very important. After investigation, the TECH Planet found that there are two main types of players in China.

Data intelligent market domestic and foreign representative manufacturers

The first category is innovative manufacturers such as Dippu Technology. Many of them have adopted a new generation of design in the technical structure, including the integration of the lake warehouse, the batch batch, and the clouds. Can meet the various analysis needs of enterprises at low cost and high performance.

"We believe that the solution of the integrated Libang one is not simply the technical capabilities of the bottom layer. It is more about meeting the final business value of the customer. Analysis of the service layer, which can provide customers with a dragon service, is also a position that is more in line with the Chinese market. "

Yang Lei mentioned that in order to better land in these technical solutions of Hucang integrated, Didopu Technology also cooperated with industry experts on the data basic platform to better use the data platform built for customers. The rigid value of the specific business is more pursuing effect services. The second type of player providing data upgrades in China is a "lake warehouse integration" solution launched by large factories based on public cloud business, such as Huawei Cloud's FusionInsight, volcanic engine LAS, Alibaba Cloud's MaxcomPute, etc.

It is the advantage of such players from the IAAS resource layer to the data layer and the application layer. Of course, many customers do not want data services to be bundled with public cloud services, which will lead to too much dependence on a platform.

Yang Lei told the TECH Planet that Ji Pu and other emerging players and public cloud platforms are also cooperative relationships. Their product FastData can also be deployed on these mainstream public cloud platforms. Customers can choose the data service of the cloud platform or independent third -party services.

In addition, the Hutu Lake Platform actually contains several parts, such as the "flow batch" data analysis and processing engine DLINK, data intelligent development platform DataFacts, which is used for corporate data scientific analysis, visual modeling, machine learning and other data scientific analysis platform DataSensense Essence

FastData real -time lake warehouse platform architecture

It is undeniable that the more agile cloud deployment is also the current market trend.

In December last year, Zhao Jiehui, the chairman and CEO of Didu Science and Technology, stated at the internal meeting of Didopu about strategic review, "The core of the strategy is not only in war (what to do), but more importantly (not to do). "In this meeting, Didu clarified the Cloud First strategy and regarded internationalization and ecology as the focus of the company's next step. This is actually the same competitive environment with the cloud platform giants as the cloud platform giants.

It is reported that within Didopa Technology, it is gradually upgrading the customer service model to provide services to multiple industries through cloud service methods. The DIC (DataInovation Center) service team from Didophara has now been upgraded to a new DIC (Data Intelligence Cloud) team to help customers quickly and cloudize deployment services.

Do not use digital "flickering" to increase business value for customers

In the process of digital upgrading of most enterprises, especially in the upgrading of large and medium -sized enterprises, commercial value has become the core assessment indicator.

In the past, enterprises focusing on the transformation of marketing applications in digital transformation. With the deepening of digital transformation, more and more enterprises have realized that optimizing the application of underlying technology in improving its own operating efficiency and supporting business scientific decisions. Enterprises pursue how to build "data houses" to bring real growth value.

This is the case with the help of Dupu Technology's FastData for digital upgrades.

Two years ago, Belle International Science and Technology Center began to cooperate with Didu Science and Technology. On the basis of the original, the data dictionary project continued to improve. It took more than 16 months to sort out nearly 600 dimensions and 1300+ indicators, and finally reached the data. The unity of logic has completed the standardization of data.

Yang Lei introduced that during the process of cooperating with Belle, based on the core capabilities of the real -time lake warehouse platform FastData, the two parties completed the unity of multiple existing warehouses in just a few months. Realize the real -time data analysis capabilities of the store level to the group level to the group level, and shorten the analysis time of the previous T+X to T+0 real -time analysis.

For example, when Belle's store manager is at work in the morning, he can see the analysis of the business data of the day after get off work yesterday. In the past, this cycle may take two to three days.

And the data analysis platform that was originally used only for the CEO service of the enterprise to various data intelligent applications, sinking to the use scenarios of front -line employees. Product services such as label factories and store manager AI assistant applications shared by the two parties can provide real -time data feedback basis and better support and nurture the work of business management.

Belle's digital construction has a benchmarking significance in the industry, because Belle not only values ​​digitalization, but has a total of thousands of technical teams in the group. It is a typical enterprise of digitalization and a partner of deep cooperation between Dupu technology.

Not only Belle, now, digital changes have penetrated into thousands of industries. More and more companies have begun to choose to cooperate with professional technical innovation manufacturers to jointly build data intelligent platforms to improve the production and operation efficiency of enterprises, further realize experience innovation, management innovation, and management innovation And model innovation.

With the deepening of digital transformation, it means that data intelligence has entered the best era. According to the "China Digital Economic Development Report (2022)" released by the China Institute of Information and Communication, the number of Chinese digital economy reached 4.55 trillion yuan in 2021, an increase of 16.2%year -on -year, accounting for 39.8%of GDP.

Of course, the trend of integration from the lake and the development process of Didopu technology proves that the "data house" is important. It is more important that the "house" that is finally created can improve the operating effect and eventually enhance the commercial value.

- END -