What's included in an enterprise-level data architecture

What is included in enterprise-level data architecture is as follows:

The data center is the core infrastructure of enterprise digital transformation. BaiLing data in many years of enterprise-level big data applications, data asset management practice realized that: data center in the enterprise digital transformation of different stages have different characteristics and operation mode, so try to summarize the proposed enterprise-level data center of the three stages of development model, share the following.

The model suggests that the development of data center will go through three stages: service-oriented, product-oriented, and ecological, which represent the three levels of data center support capability, application mode, and value embodiment step by step.

Note that these three phases are not distinct: many companies tend to exhibit most of the characteristics of one phase, with some localized characteristics of the other two phases, but this does not prevent us from independently studying these three phases from a macro perspective.

In addition, this model is based on the observation of traditional large state-owned enterprises (e.g., energy and power, operators, manufacturing, finance, etc.) and other "non-native digital enterprises," and may not be applicable to Internet e-commerce, online entertainment, and other "native digital enterprises. "

Service-based

This is the first stage in the development of data middleware. In this stage, the users (business departments) of the data middle office tend to use direct data consumption (e.g., querying databases directly through SQL) or microservices to obtain data, process it into metrics, and form reports, kanban boards, and other applications, which are used to understand the current state of the enterprise's business, as well as to evaluate the performance of the departments and employees.

At this stage, the underlying layer of the data middleware is generally basic data services such as ETL, scheduling and monitoring, data governance, etc., and the data architecture of the enterprise tends to be a vertical (chimney-type) departmental data architecture, with the problem of data silos being more obvious.

While the data middleware operations team tries to horizontally integrate data in an enterprise-level thematic way, this integration is partial, fragmented, and inefficient because the data model does not y consider the needs of enterprise-level horizontal applications in the very first design phase. Early pain points in this phase of work tend to be poor data quality, difficult integration of data models, and unstable data platforms/tools.