Foxconn Cloud Platform and Industrial Internet Practices: Microservices, DevOps, and Cloud‑Native Architecture

I am pleased to share some practices of Foxconn Cloud today; I am responsible for the Foxconn Cloud platform and industrial‑internet‑related products and solutions. First, we will discuss why microservices, DevOps, and cloud computing have become pervasive across industries; second, we will explore the unique design ideas of microservice architecture and how it differs from traditional SOA; finally, we will look at practical applications of cloud platforms and microservices in the industrial domain.

-- From Ford’s Model T --

The picture is from Henry Ford’s 1913 Model T. Those cars were assembled step‑by‑step from steel plates on a production line. Today, companies such as Uber and Tesla have changed travel habits, and in industrial manufacturing a large number of intelligent robots are replacing manual labor.

Industrial Internet is gradually reshaping the industry landscape. From Ford to Toyota to Elon Musk, each industrial era has its distinct characteristics. Ford reduced cost through scale, Toyota’s Taiichi Ohno introduced lean and just‑in‑time production after World War II, and Musk’s lean approach has greatly boosted Tesla.

Process control, lean manufacturing, and the waterfall model of the 1990s (CMM) were linear and incremental. With the rise of cloud computing, agile development, DevOps, and microservices, these new technologies have dramatically transformed our business, ushering us into a software‑defined world.

Agile development overturns the traditional waterfall model by enabling rapid iterations, allowing developers to quickly prototype and evolve products. DevOps culture, processes, and tools break down silos between development and operations, enabling continuous delivery pipelines, gray‑release strategies, and production‑grade deployments.

-- Characteristics and Design Ideas of Cloud‑Era Application Architecture --

What does a cloud‑native application change? It changes the way we design and think about traditional monolithic software architectures.

Why adopt this approach? Traditional architectures have many advantages, especially the service‑oriented design and deployment models that dominated from the 1980s to the 1990s, including mainframe applications in banking.

With the evolution of the Internet, distributed services, and cloud platforms, microservices emerged. Companies like Amazon and Google design their systems using microservices.

How do we run such applications? Cloud has evolved from traditional data centers to on‑demand public, private, and hybrid clouds, including server‑level implementations.

These advances enable 24/7/365 service availability, essential for manufacturing and banking applications, while supporting industrial‑grade workloads and business agility without downtime.

How to develop projects in practice? Consider Conway’s Law: the architecture of a system mirrors the organization’s structure. Time is limited, but a project can always be completed with proper planning.

The twelve‑factor methodology provides good guidance for building cloud‑native services, helping us evaluate whether our design, development, and deployment processes are sound.

The cube model illustrates common service‑splitting and design principles. The X‑axis focuses on horizontal scaling of data and services, requiring service registration, stateful/stateless separation, and load‑balancing. The Y‑axis emphasizes functional decomposition, often resulting in hexagonal microservice shapes that enable small, autonomous teams.

Horizontal scaling relies on front‑back separation and decentralization. Microservices differ from traditional SOA by avoiding centralization; they adopt stateless/stateful patterns and expose services via ports, similar to MVC transformations.

Vertical division is achieved through functional decomposition—extracting core order services, for example—used widely in internet, finance, and manufacturing sectors, with data partitioning and high‑availability designs.

We adopt Domain‑Driven Design (DDD) to build applications. DDD, though long‑standing, focuses on business‑centric modeling, separating infrastructure (persistence), core domain logic, and user interfaces, facilitating horizontal scaling and clear responsibility boundaries.

-- Industrial Internet Practice --

Many ask about the difference between Industrial Internet and IoT. The Industrial Internet builds on a universal internet platform, leveraging cloud, mobile, and 5G for B2B scenarios, and adds large‑scale data analysis and domain‑specific models.

Our project includes an overview architecture with an edge layer. Industrial Internet is challenging because industrial‑grade products are scarce; for example, GE’s Predix supports energy, healthcare, and aerospace.

Edge computing and data acquisition are critical. Industrial data volumes are massive; processing at the edge reduces bandwidth, latency, and cost before sending to the cloud.

Both public and private clouds are viable; we focus on the upper‑layer applications, adopting multi‑cloud strategies, PaaS platforms, and SaaS services for robotics, SMT, AGV, etc.

We need a developer platform that provides pipelines to package, publish, and deploy industrial microservices, supporting DevOps and fostering an ecosystem for the Industrial Internet.

Our Foxconn cloud ecosystem also incorporates AI algorithms for industrial models, demonstrating end‑to‑end solutions.

Organizational structure directly impacts a company’s capability to adopt microservices. Front‑end, product design, testing, operations, and support must be aligned, and flattening the organization enables decentralization.

Setting OKR goals is essential; each team must know its objectives to drive industrial‑project platforms.

Technology selection leans toward Cloud‑Native solutions; we aim for true decoupling, transitioning from monoliths to microservices and DevOps to ensure long‑term business growth.

We redefine failure: it is not scary, but how we define and learn from it matters. In the industrial internet, rapid, low‑cost experimentation and fast prototyping are crucial.

The success factors of microservice projects form a golden triangle: customer‑centricity, microservice architecture design, agile organization, and DevOps pipelines.

China DevOps Community Online Sharing

This Tuesday at 8 pm

Scan the QR code now to discuss microservice design with Professor Xu Zhouping.