How Distributed Architecture Is Powering the Next Wave of Chinese Bank IT

Two Approaches to Domesticating Bank Systems

Industry consensus identifies two ways to achieve domestic‑focused banking IT: (1) dissect and replicate the principles and architecture of foreign products, then rebuild a comparable system using local technology—such as creating a home‑grown mainframe‑class platform or a domestic database to replace Oracle; (2) adopt a distributed model that substitutes traditional mainframes with clusters of ordinary x86 servers.

Current Domestic Progress

In storage, mature domestic products can already replace EMC solutions. In the database sector, many Chinese vendors are improving, yet performance still lags behind Oracle. Distributed application technology is maturing and is expected to gradually replace IBM mainframes.

Distributed Architecture Characteristics

The distributed architecture is built on x86 and cloud computing, featuring data sharding and read/write separation. It scales horizontally by adding servers, each acting as an independent node whose failure does not affect overall availability.

Evolution of Architecture

The shift has moved from monolithic to vertical silo, then to SOA, and finally to micro‑services. Compared with centralized monolithic systems, distributed designs disperse functionality across many micro‑services.

Distributed vs. Centralized: Comparative Advantages

Economics: Centralized mainframe solutions often cost billions, while x86‑based distributed clusters are far cheaper.

Scalability: Mainframes improve performance by enhancing a single machine, which has limits; distributed systems scale by adding more nodes.

Autonomy: Mainframe reliance on foreign vendors reduces autonomy; x86 clusters increase self‑reliance.

Flexibility: Centralized systems are closed, whereas distributed platforms draw from open‑source ecosystems, offering greater flexibility.

Operations Difficulty: Distributed clusters involve many servers, making maintenance harder; mainframes have fewer components and simpler ops.

Availability: Distributed infrastructures face challenges in cloud platforms, replication, and synchronization, but generally achieve better availability than centralized setups.

Consistency: Centralized architectures provide strong consistency; distributed systems rely on multi‑node coordination to achieve high availability.

Digital Currency’s Influence on Bank IT

Digital currency adds new IT construction demand across three dimensions:

Issuance side: The central bank will build its own digital‑currency system, offering limited opportunities for commercial IT vendors.

Operation side: State‑owned large banks will need extensive digital‑currency transaction, supervision, and reporting systems, typically sourced from long‑term partners; smaller banks will also depend on IT vendors for integration.

Circulation side: ATMs and POS terminals must be upgraded to handle digital‑currency transactions alongside traditional cash.

Market Demand and Service Segments

Rising bank‑IT optimism drives comprehensive demand, mainly focused on:

IT solutions: Channel solutions for customer outreach, core business systems (payments, credit, middle‑office), and management/risk/compliance tools.

Outsourcing services: Development, cash‑management, call‑center, data‑management, and integrated‑business services, especially as large banks face staffing pressures and upcoming retirements.

Integration services: Though lower margin, integration deepens vendor‑client relationships.

Large banks prioritize core‑system upgrades, while smaller banks gradually adopt distributed architectures for channel and risk‑management modules.

Competitive Landscape

The new generation system, developed for China Construction Bank, targets large banks; thus, Jianxin Jinke (建信金科) is likely to win more large‑bank contracts in the short term. In the long run, it may expand to mid‑size banks. Changliang Technology and Shenzhou Information, whose core business focuses on mid‑size banks, are scaling up and attempting to penetrate large‑bank markets, leading to eventual direct competition.

Fusion Architecture and Performance Metrics

The next‑generation system combines distributed and centralized layers (7+1 layers, 12 platforms), reducing reliance on foreign vendors and achieving over 90% in‑house development. Key statistics after deployment at China Construction Bank include:

More than 83% of applications run on x86 cloud.

Host utilization dropped from 90% to 70% while transaction volume grew 33%, keeping host resource usage flat.

This hybrid approach leverages the security and stability of mainframes with the cost‑effectiveness and scalability of distributed platforms.