On January 4, 2023, in order to further enable the release of the value of data elements and strengthen exchanges and cooperation in the data asset industry, under the guidance of the China Academy of Information and Communications Technology and the China Communications Standards Association, the China Communications Standards Association Big Data Technology Standards Promotion Committee (CCSA TC601) The fifth Data Asset Management Conference hosted by the company was held online.
At the meeting, Du Xiaozheng, general manager of the Business Analysis Division of CEC Financial and chairman of the Data R&D Committee of CEC Financial, gave a keynote speech, analyzing the development trends of data management, management hot spots and pain points of financial customers, and introduced Yuanqi. The data asset management platform, "China Electronics Financial Trust Data Governance White Paper", "1+2+3" data asset consulting system, etc. also elaborated on the "Yuanqi" as the cornerstone of China Electronics Financial Trust in the field of financial big data. Driven by the concept of continuous innovation, we continue to be committed to assisting financial institutions to achieve rapid digital transformation at the business level and open up the entire process of business digitization and data commercialization.
General Manager of CEC Jinxin Business Analysis Division
Du Xiaozheng, Chairman of China Electronics Jinxin Data R&D Committee:
1. Development history and technology trends of data management in the financial industry
■ Development history: Data management in the financial industry has entered a new era of data capitalization
Since around 2000, the financial industry has begun to do related data management. From the earliest data warehouse to data governance, data applications, and later big data, until today data management has entered a new era. We summarize it as the era of data assetization , or the era of localization .
Why two names?
The reason why it is called the era of data capitalization is because the country has promoted data as a factor of production and juxtaposed data with other factors of production. Therefore, we say this is an era of data capitalization; it is called the era of localization because in 2018 After that, the trend of localization of data management was very obvious, so in the new era, data management is both an era of data assetization and an era of localization.
This era has three outstanding characteristics:
First, start a large-scale migration to domestic data platforms;
Second, data asset assessment and related data asset inventory have become one of the key tasks of various financial institutions;
Third, so far, data governance and data analysis have entered a new trend. The entire data governance system, from standards to tools, has become regulatory scenario-driven.
■ Data management technology trends
When talking about the technology trends and characteristics of data management, you can refer to the 2022 maturity curve released by Gartner on the left side of the figure above. It can be seen that there are some important technology trends, such as Lakehouse, DataOps, Cloud Data Ecosystem, Data Fabric, and Active Metadata Management. Although many of these new technologies are summarized from overseas or European and American technology application trends, they are also applicable to domestic situations, but the time may be delayed by 1-2 years compared with Europe and the United States.
Looking at the general trend, there are the following development directions:
■ 1. Cloudification of data management technology
In recent years, cloud native has been a very important trend. Data management technology has also gradually become cloud-based, and has transformed from a relatively independent and isolated state in the past to the current trend of technology upgrades based on enterprise cloud platforms.
■ 2. Deep binding of data management and AI
We used to think that data management is data management, AI is AI, and there is a difference between the two. But today, when we talk about AI, we will definitely mention data management, and data management, as an important application, will definitely mention AI when we talk about it. These two technologies are actually a deeply bound and mutually reinforcing relationship. AI has also gradually transformed from the previous Model-Centric to Data-Centric.
■ 3. Centralization of data capabilities
The need for cross-platform management and control has given rise to the need for business or technical departments to quickly access and iterate data, which has given rise to the need for DataOps.
■ 4. Data platform self-evolution
In recent years, the data platform has begun to gradually evolve to the third-generation data platform, and Gartner also defines the data platform of the Data Fabric architecture as a new generation of data platform. The biggest feature of this architecture technology platform is that it has the function of self-evolution and self-evolution, that is to say, the future data platform will not be built gradually from scratch based on our traditional method, based on data-driven and technology-driven methods, but based on the application The driving method uses active metadata technology and knowledge graph technology to gradually improve it according to the access situation, and finally forms the self-evolution process of the data platform.
The above is the entire development trend of data management that we have observed in recent years.
2. Hot spots of financial customer data management
What are the current data management hot spots for customers in the financial field? We summarized and decomposed into four directions.
■ Hot spot one: organizational structure
After the People's Bank of China and the China Banking and Insurance Regulatory Commission respectively issued their own relevant digital transformation regulations, each bank has also accelerated its own digital transformation process. The most obvious feature is the adjustment of each bank's organizational structure.
From the end of last year to the beginning of this year, large state-owned banks and joint-stock banks have established their own first-level data management departments. This means that each bank directly promotes the data departments that used to be in technology or under some business departments to the first-level department. big adjustment. The deep-seated reason behind the adjustment is actually to hope that there is a first-level department that can promote the digital transformation of the entire bank across teams, and realize the process of transforming the bank's previous data management from a cost center to a profit center.
■ Hot spot 2: technical architecture
We learned from the customer's situation that everyone is doing their own migration process this year, but the overall localization migration is not a simple migration from platform to platform, and is often accompanied by the migration and evolution of the entire architecture.
During the migration process, we found that the integrated architecture of lakes and warehouses is a hot spot that many customers are currently admiring. Many customers choose to upgrade the original data platform architecture as a whole when migrating projects, instead of only supporting analysis. The mode is going in the direction of supporting both natural data and complex processing analysis, fully supporting the integration of lakes and warehouses.
In addition, financial institutions have also begun to embrace open source related technologies in large quantities. For databases such as ClickHouse and Doris, which are widely used on the Internet, financial customers have begun to gradually promote them in certain scenarios.
■ Hot Topic 3: Data Assets
There are two hot spots for data assets:
The first hot spot is the evaluation of data assets, which is also a hot spot that various industries are now exploring. In 2021, China Everbright Bank and Shanghai Pudong Development Bank released their own white papers respectively; at the end of 2022, China Everbright Bank further released its own white paper on data asset assessment. It can be said that China Everbright Bank has become a pioneer in practice among its peers. We believe that these are of great reference significance to peers. Customers in the same industry are also gradually exploring and improving their own data assets, and conducting relevant data asset inventory and evaluation.
The second hot spot is the management of data assets. In traditional data management, data assets are often managed by middle and back-end departments. But today, we see that enterprise architecture modeling and data platform models are gradually being connected, and the data management department is gradually moving from the previous middle and back-end departments to the front-end department. At the same time, financial customers will consider the design of the data model when modeling front-end enterprise architecture, so that a set of enterprise-level models supports both application models and data management models.
■ Hot Topic 4: Data Security
The "Data Security Law" was released in 2021 and stipulated many requirements for data security classification and classification. With the emergence of some new technologies, such as privacy computing, it also solves the problem of data availability and invisibility to a certain extent, which promotes financial institutions to fully promote the implementation of data security solutions and the development of related applications.
3. Pain points of financial customer data management
Earlier we talked about the development history, technology trends and customer hotspots of the entire data management. In addition, many financial customers also have a lot of pain points, which are actually very important inputs for us.
We simply summarize it into the following four directions:
■ Pain point 1: Decentralized data management
As we all know, banks’ data management systems are gradually established, and most bank data are scattered in different data management departments. However, the first-level management department is currently established to manage the bank's data management platform and data assets. At this time, it will face an arduous task - the data management department lacks enough energy to manage the entire platform.
■ Pain point 2: Diversified data needs
The diversification of data requirements has led to complex management methods for data requirements. Data requirements often lack the support of a continuously iterative matching system. How to continuously iterate subsequent data requirements and how to quickly build a new platform based on the diversification of data requirements so that data applications can achieve expected results is a big challenge.
■ Pain point three: Fragmentation of data governance
As of today, data governance has reached the third wave. The first wave is mainly about data standards, the second wave is about data products, and the third wave is about data governance methods driven by supervision or other scenarios. At the same time, we can also see that data governance can no longer simply follow the approach of "treating headaches and treating painful feet" in the middle and back offices. We need to comprehensively promote data governance from the source.
■ Pain point 4: Accurate assessment of data value
As data is a factor of production, in order to realize relevant data capitalization and trade data assets, the premise is to carry out inventory and valuation of data assets. How to carry out an accurate valuation? How can we come up with a fair value? This is yet another challenge facing financial customer data management. (Unfinished)