Trend 1: The shift from traditional centralized to distributed middleware is now a consensus
Traditional enterprises widely use centralized architectures. The IOE architecture is a prominent example. A traditional centralized architecture typically consists of three main components: powerful minicomputers with high single-point reliability, stable and expensive storage, and a centralized database. Heavyweight Java 2 Platform, Enterprise Edition (J2EE) containers, such as Websphere, or transaction middleware such as CIS, serve as the runtime to support business applications and execute business logic. However, as business models change and the need for digital transformation grows, the shortcomings of traditional centralized architectures have become apparent:
Complex monolithic logic: All modules are bundled into a single package that can be several gigabytes in size. This leads to slow compilation, slow startup, code conflicts, and difficult releases.
Poor scalability: Significant architectural bottlenecks prevent scaling, which makes it difficult to meet growing business demands.
High cost: The architecture requires specialized, high-end, and expensive hardware.
As technology evolved, distributed architectures emerged to address these pain points. The move toward distributed middleware is now widely accepted. Internet companies in China have widely adopted distributed architecture technologies, with many successful cases. In traditional industries, especially finance, there is also a consensus to transition to distributed systems. For example, the Industrial and Commercial Bank of China (ICBC) began its IT architecture transformation in 2014, moving from a primarily centralized system to one that integrates both centralized and distributed architectures.
Trend 2: The rise of cloud-native concepts drives product innovation
Cloud-native is an approach to building and running applications that takes full advantage of cloud computing. Cloud-native applications can be built, released, and deployed quickly and frequently. They are decoupled from the underlying hardware and operating systems. This makes it easy to meet scalability, availability, and portability requirements while being more cost-effective. This approach also makes organizations more agile by enabling the formation of small, functional teams. It fosters better integration of people, processes, and tools, and enables closer collaboration between developers, testers, and O&M staff.
The cloud-native concept was introduced in 2015. In June 2018, the Cloud Native Computing Foundation (CNCF), initiated by Google, officially released version 1.0 of the cloud-native definition: "Cloud-native technologies empower organizations to build and run scalable applications in modern, dynamic environments such as public, private, and hybrid clouds. Representative technologies include containers, service mesh, microservices, immutable infrastructure, and declarative APIs."
Cloud-native provides value across the entire application lifecycle, from development and testing to O&M. This has driven many innovations in middleware, such as containerization, microservices, service mesh, and serverless computing. These innovations help build loosely coupled systems that are fault-tolerant, easy to manage, and simple to monitor. Combined with reliable automation, cloud-native technologies enable engineers to make frequent and predictable major changes to systems with ease.
Trend 3: Open source middleware is recognized but faces production pressure
Many open source middleware products, such as Dubbo, are now widely used. However, their users are primarily Internet companies. When used in production for enterprise-grade and financial-grade applications, open source products often fall short in several key areas:
No Service-Level Agreement (SLA) guarantees.
Lack of professional support services.
Difficulty meeting basic enterprise requirements for high availability (HA) and disaster recovery.
Insufficient productization and a lack of enterprise-grade management and control capabilities.
Traditional enterprises have a limited talent pool for open source distributed architecture technologies.
Therefore, many traditional enterprises are exploring a hybrid approach. They combine open source and commercial solutions to implement distributed architectures while retaining independent control.
Trend 4: Chinese technology competes with global players
China has a massive population and a world-leading Internet industry. As a result, Chinese Internet companies have turned their practical experience into products. These companies have handled enormous traffic bursts from massive user bases across diverse application scenarios. After years of refinement, these products are now gaining international influence.
As early as 2012, Alibaba open-sourced Dubbo. It has become a popular microservice framework and was contributed to the Apache Software Foundation.
Several open source projects from Ant Group's SOFAStack have been added to the CNCF Cloud Native Landscape. For example, Mosn, a Service Mesh data plane agent written in Go, has attracted widespread attention.
Seata, jointly open-sourced by Alibaba and Ant Group, has become the most popular distributed transaction framework on GitHub.
Therefore, for Chinese enterprises that are upgrading their middleware, domestic distributed middleware, which offers advanced technology and a rich ecosystem, is an excellent and practical choice.