Data Middle Platform: Vision, Architecture, and Business Value

This article summarizes TVP expert Shi Kai’s live lecture on the concept, purpose, and essence of the Data Middle Platform (also called Data Mid‑Platform), which is regarded as the next stage of big‑data development and the focus of industry discussion in 2019.

Definition and Vision : The Data Middle Platform is a platform that provides data services/products to all data consumers in an enterprise, turning raw data into business‑driven value. It aims to make data closer to business, enable fast data service development, and support digital transformation.

Why It Matters : Traditional data warehouses and Business Intelligence (BI) tools serve mainly reporting needs for decision makers. They are not designed to serve front‑end business systems directly. The Data Middle Platform bridges this gap by offering real‑time, reusable data APIs that can be consumed by applications, reducing inefficiencies, inconsistencies, and skill gaps between application developers and data engineers.

Key Business Drivers (ranked by a survey of 460 respondents): Closer proximity of data to business Provision of data services Direct business value generation Rapid development of data services Additional expectations include multi‑scenario support, unified data, and breaking data silos.

Comparison with Related Concepts : Data Warehouse – a technology‑specific storage system with standardized products. Data Platform – a broader technical foundation for collecting and processing large‑scale data. BI – primarily visual reporting for decision makers. The Data Middle Platform differs by focusing on service‑oriented data delivery to business front‑ends.

Typical Architecture Examples (illustrated with Alibaba, Cainiao, Suning, Didi, OPPO, and telecom cases): Alibaba – One Data, One Entity, One Service; unified data entities and services across retail. Cainiao – Service layer, data layer, and management suite. Suning – Transition from process‑driven to data‑driven architecture. Didi – Emphasizes data culture, intelligent data catalog, and agile governance. OPPO – Four‑layer model: tool layer, data warehouse core, data market, and data products. Telecom – Centralized data services for various business lines.

Six‑Capability Model of a Data Middle Platform : Data asset planning and governance Data acquisition and storage Data sharing and collaboration Business value discovery Data service construction and governance Data service measurement and operation This model supports the overarching goal of building a data‑driven intelligent enterprise.

Practical Takeaways : Even without an existing data platform or warehouse, enterprises can start by identifying business data needs, exposing them as APIs, and iteratively building data services. The focus should be on delivering fast, secure, and reusable data products that directly empower business processes.

Overall, the lecture emphasizes that the Data Middle Platform is not a single technology but a combination of tools, methods, and organizational practices that enable digital transformation through data.