System Architecture with Isolated Data Islands
Multiple sources and heterogeneous of the underlying computing and storage architecture result in system silo and isolated data island
In the early business development period, companies would deploy IT systems based on vertical and personalized business logic to solve business problems. Most of the various information systems were purchased and built independently. They are deeply coupled with processes and underlying systems and have many cross-correlations with upstream and downstream systems, resulting in multiple isolated data islands in the company and huge difficulty in achieving total interconnection of information. Meanwhile, in the exploration of new platforms, new businesses, and new markets, the system can’t be directly reused or quickly iterated. The data generated in the process can’t be interconnected with data accumulated in the traditional model, which exacerbates the problem of isolated data islands. Since it is hard for scattered data to deal with the changes of front-end businesses, support corporate operation decisions, there is an urgent need for a mechanism that combines the old and new models and integrates the data scattered on various islands to gain data service capabilities.
The Connotation of the Data Middle Platform
Data middle platform is not only a technical concept, but also a corporate management concept
Data middle platform is in the middle of the front end and the back end. It is an enterprise-level data sharing and ability reuse platform. It contains a series of data components or modules. Through the unified collection, processing, storage, calculation, analysis, and visualization of an enterprise’s massive multi-source and heterogeneous data, the data middle platform transform data into a company’s core asset, providing data resource and capability support for the business front end to achieve refined operation driven by data. Based on its information construction foundation and business characteristics, the enterprise defines the capabilities of the data middle platform and selects and uses data components to build a middle platform. In a broad sense, the data middle platform is a model and concept of the organization and management of an enterprise, integrating a company’s strategic determination, organizational structure, and technical structure. The enterprise strategically builds a unified collaborative base, that is, a middle platform organization, to coordinate and support various business departments, expand business boundaries by technology, and provide room for the growth of new businesses and new departments.
System implementation requires the multi-dimensional capabilities of both the supply and demand parties
The construction of a data middle platform involves many technologies. In the aspect of technology structure, it needs scalability, agility, lightweight, emphasizes interaction with the front-end, and realizes application functions through service arrangement to meet the demand of the front end. Data middle platforms now follow the design principle of “high cohesion, loose coupling, integrate distributed model, micro service, container cloud, DevOps, big data processing and architecture with high availability, high performance and high concurrency and have formed a mature methodology. Currently, the difficulties faced by data middle platform construction are mainly how to combine mature technical solutions with real conditions and characteristics of the industry and enterprises and make feasible solutions for data middle platform construction based on application scenarios. The enterprise’s resource allocation ability, management experience, organizational structure, business organizing ability, as well as data middle platform service providers providing consulting and planning services for corporate data governance in enterprise middle platform construction, gradually become key elements in the construction of data middle platforms.