09 Middleware - What Exactly to Share After Digital Transformation #

Hello, I’m Ouchuangxin.

In the previous lecture, we learned about the design thinking behind the layered architecture and mentioned that this design thinking is very beneficial for center stage construction. So today, let’s talk about Center Stage.

Center Stage is a hot topic in digital transformation. After Alibaba proposed the concept of Center Stage, many others have come up with various interpretations. Today, we will mainly discuss business center stage and data center stage. As part of the overall transformation of enterprise digital center stage, I will also briefly discuss the design thinking behind the front end and back end.

Many companies actually have years of experience in building large platforms, so when Center Stage started to gain attention, I believe you must have heard many skeptical voices. For example, some people say, “Center Stage is just a fancy term, isn’t it just another term for platforms that have been around for many years?” Indeed, Center Stage originated from platforms, but its strategic significance is much higher than that of traditional platforms.

After completing this lecture, you will have a clear understanding of the differences between platforms and Center Stage. What exactly is Center Stage? Should the construction approach of traditional companies’ Center Stage be the same as Alibaba’s…

Is the platform really a mid-end? #

After Alibaba proposed the mid-end strategy, many companies began to benchmark their own systems against Alibaba’s mid-end. Some companies had already completed the split of their centralized systems into a unified platform more than ten years ago, achieving the evolution from traditional monolithic applications to large platforms. They separated the common capabilities and core capabilities, solving the problem of redundant investment and construction of common modules.

So is this what Alibaba calls the mid-end? Before answering this question, let’s first understand what Alibaba’s mid-end is.

The predecessor of Alibaba’s business mid-end was the shared platform, which was more like a resource team. They undertook the requirements of various business units and customized development for the business units on the basic services. The goal of Alibaba’s business mid-end is to treat the core service chain (membership, goods, transactions, marketing, stores, fund settlement, etc.) as a platform product as a whole, providing business solutions for the front-end business instead of independent systems.

Now let’s analyze the differences between the traditional enterprise platform strategy and Alibaba’s mid-end strategy.

The platform only separates some of the common capabilities as a shared platform. Although it can provide common shared services to the outside through APIs or data to solve the problem of redundant system development, this type of platform has not achieved comprehensive integration of pages, business processes, and data from the front-end to the back-end with other platforms or applications within the enterprise, and the core business service chains have not been considered as a holistic solution. The platforms are still separate and independent.

While the platform solves the problem of reusing common capabilities, it still has a long way to go to achieve the goals of the mid-end!

What is the Middle Platform? #

“There are a thousand Hamlets for a thousand readers,” which perfectly describes the various definitions of the middle platform in the tech industry. There are many different perspectives.

Let’s start with Alibaba’s definition of the middle platform: “The middle platform is a fundamental concept and architecture. We use the middle platform approach to build and connect all basic services, providing support for upper-level business. The business middle platform primarily supports online business, while the data middle platform provides basic data processing capabilities and many data products for all business units to use. The business middle platform, data middle platform, algorithm middle platform, and so on collectively provide support for upper-level business.”

Now let’s look at ThoughtWorks’ definition of the middle platform: “The middle platform is an enterprise-level capability reuse platform.”

In summary, we can extract several key words related to the middle platform: sharing, connectivity, integration, and innovation. Connectivity refers to the connection between the front-end and middle platform, integration refers to the integration of front-end processes and data, and sharing supports the development and innovation of front-line business in a collective manner.

In my opinion, the middle platform first and foremost embodies an enterprise-level capability. It provides an enterprise-level holistic solution, addressing the issues of capability sharing, connectivity, and integration for small enterprises, large corporations, and even ecosystems, supporting business and business model innovation. By connecting platforms and integrating data, a consistent user experience can be provided, supporting front-line business in a more agile manner.

The middle platform originated from platforms, but compared to platforms, it primarily reflects a shift in mindset. It mainly manifests in three key capabilities: rapid response to front-end business, enterprise-level reusability, and seamless connectivity and integration of design, development, page operations, process services, and data from front-end to middle platform and back-end.

Among these, the most critical are the rapid response capability and enterprise-level seamless connectivity and integration, especially for large-scale enterprises with cross-industry operations.

What should be shared in the digital transformation center? #

Compared to internet companies, traditional companies have more diversified channel applications, including in-store applications for internal staff, internet e-commerce applications for external users, and mobile app applications. These applications may target different users and scenarios, but they have similar functionalities that essentially cover core business capabilities. In addition, traditional companies may also open up some core application pages or API service capabilities to third parties in the ecosystem to mutually support development.

In order to adapt to the development of different businesses and channels, many companies in the past developed many independent applications or apps. However, due to the lack of enterprise-level overall planning during the initial IT system construction, the integration between platforms was not well done, resulting in poor user experience. Most importantly, users don’t want to download so many apps.

To improve the user experience and achieve unified operations, many companies have started to reduce the number of apps and integrate all capabilities within the enterprise through one app, connecting all core business processes on the front end.

Since the business models and the development of IT systems in traditional companies are not exactly the same as internet companies, the strategies for constructing a central platform in traditional companies should also have some differences from Alibaba’s central platform strategy, and the content that needs to be shared is also different.

Due to the diversification of channels, traditional companies not only need to centralize common capabilities to achieve the accumulation, sharing, and reuse of common capabilities, which corresponds to the generic domain or support domain of DDD (Domain-Driven Design), but they also need to centralize core capabilities to meet the demand for sharing and reuse of core business capabilities across different channels, avoiding the problem of “dual backend and dual frontend” of traditional core and internet applications. Here, core capabilities correspond to the core domain of DDD.

This falls under the scope of business center. We need to solve the problem of connecting core business processes and sharing services across different channels. In addition, we also need to solve issues such as data fragmentation, data integration, and business innovation after system microservices are decoupled. This falls under the scope of data center, especially when we adopt a distributed architecture, we should pay more attention to data integration and sharing issues after microservices decoupling.

In summary, when designing and planning a central platform, we need to consider the collaboration of front-end, center, and back-end applications within the enterprise as a whole, to achieve the sharing of front-end pages, processes, and services across different channel applications, as well as the connection of core business processes and the integration and sharing of front-end processes and data, supporting business and business model innovation.

How to Achieve Collaboration between Front, Middle, and Back Offices? #

The ability of enterprise-level collaboration often reflects the ability to collaborate between the front, middle, and back offices.

If the business middle office is compared to specialized branches of the army, such as the army, rocket forces, and air force, it mainly plays a tactical professional role. The front office is the combat force, which needs to dispatch the capabilities of the business middle office according to the demands of the front-line battlefield, achieving capability integration and maximizing efficiency. The data middle office is the information intelligence center and the joint command headquarters, which can gather various data, complete analysis, and develop strategic and tactical plans. The back office is the logistics force, providing technical support. Let’s discuss each of them below.

1. Front Office #

Many early systems of traditional enterprises were built based on business domains or organizational structures. Each system had its own front end, which operated independently. User operations were vertical, requiring logging into multiple systems to complete the entire business process.

The construction of the front office after the middle office should have a comprehensive solution that considers business boundaries, processes, and platforms to achieve connectivity and integration of different middle office front-end operations, processes, and interfaces. No matter how many middle offices there are in the back end, users will only feel that there is one front office.

In front office design, we can draw inspiration from the design principles of micro frontends, not only achieving front-end decoupling and reuse within the enterprise but also integrating front office business through dynamic composition and process arrangement of micro frontend pages based on core links and business processes.

Front-end pages can naturally integrate into different terminal and channel application core business links, achieving front-end page, process, and feature reuse.

2. Middle Office #

Most of the core businesses of traditional enterprises are developed based on a centralized architecture, and single-system applications have poor scalability and elastic scalability, making them unable to adapt to fluctuating Internet business scenarios. Most data applications extract data through ETL tools to achieve data modeling, statistics, and report analysis. However, due to the lack of data timeliness and integration capabilities, and the fact that traditional data applications were not originally designed for the front end, it is difficult to respond quickly to front-line business needs.

The construction of the business middle office can adopt the domain-driven design method, extracting reusable common capabilities from various monolithic systems through domain modeling, adding them together, and using the microservices architecture pattern to build a shared common capability middle office.

Similarly, we can use the microservices architecture pattern to build a reusable core capability middle office that is oriented towards different channels and scenarios. The business middle office provides API services to the front office, third-party systems, and other middle offices, achieving the reuse of common capabilities and core capabilities.

However, you need to remember that during the process of breaking down traditional centralized monolithic systems into microservices based on business responsibilities and capabilities and building the middle office, more and more independently deployed microservices will be created. Although this enhances the application’s elasticity and high availability, due to the physical isolation of microservices, some original internal system calls will become cross-microservice calls. Combined with front-end and back-end separation, microservices decomposition will further separate data, increasing the difficulty of integrating enterprise-level applications.

Without appropriate design and guiding principles, if the relationship between the front office, middle office, and back office is not properly handled, it will further aggravate the isolation and fragmentation of front-office processes and data. The main goal of a data center is to bridge data silos, achieve business integration and innovation, including three main functions:

First, complete the collection and storage of enterprise-wide data, realize the aggregation and centralized management of middle-stage data in different business categories.

Second, process and handle data according to standard data specifications or data models, and form data applications for different themes and scenarios, such as customer views, agent views, channel views, and institutional views.

Third, establish a business-driven data system, deeply extract data value based on various dimensions of data, and support business and business model innovation.

Correspondingly, the construction of the data center can be divided into three steps:

The first step is to achieve the aggregation of middle-stage business data, solve the problems of data silos and primary data sharing.

The second step is to achieve deep integration, processing and sharing of enterprise-level real-time or non-real-time all-dimensional data.

The third step is to extract data value, support business innovation, and accelerate the process of converting data into business value.

The data center is not limited to analytical scenarios, but also applicable to transactional scenarios. It can be built on top of data warehouses or data platforms and provided to business systems after data is made service-oriented. Based on the technology of capturing database logs, the timeliness of data is greatly improved, which provides good support for transactional scenarios.

In summary, the data center mainly completes the integration and processing of data, extracts the business value of data, supports business innovation, and provides data sharing services to the outside world.

3. Backstage #

When people mention the middle stage, they naturally ask: “Since there is a front stage and a middle stage, is there a backstage, and what are its responsibilities?”

Let’s take a look at Alibaba’s positioning of the front stage, middle stage, and backstage.

The front stage is mainly aimed at customers and terminal sellers, achieving marketing promotion and transaction conversion; the middle stage is mainly aimed at operational personnel, completing operational support; the backstage is mainly aimed at backstage management personnel, achieving process review, internal management, and logistical support, such as procurement, manpower, finance, and OA systems.

For the backstage, in order to meet internal management requirements, many people tend to embed these management requirements into core business processes. Generally speaking, this type of internal control management requirement has high requirements for authority, control rules, and processes, but most management personnel only participate in the review of a certain local business link. These complex management requirements will artificially increase the difficulty of integrating front-end interfaces and core processes of different channel applications, as well as the complexity of software development.

When designing process review and management functions, we can consider aggregating functions according to roles or positions, separating complex management requirements from common core business processes, and referring to the construction mode of mini-programs, embedding them into front-end apps or applications through specific program entry points.

After the management requirements are removed from the front-end core business process, the front-end application will have better versatility. It can more easily achieve the integration of front-end interfaces and processes of various channels. A front-end application or app can serve external Internet users and internal business personnel in the same way, thereby promoting the integration of traditional channel and Internet channel application front-ends.

Summary #

Today we mainly discussed some ideas about the construction of the middle platform. The transformation of an enterprise’s middle platform is not just the work of the middle platform itself, we need to consider the collaboration, sharing, integration, and connectivity of the front-end, middle platform, and back-end as a whole.

The front-end achieves front-end integration between different channel applications through page and process sharing, while the middle platform achieves service sharing through APIs. The integration of front-end, business middle platform, and data middle platform can realize the integration of traditional applications and internet applications, thus solving the problem of “dual back-end cores and dual front-end skins.” Only when capabilities are reused and front-end processes and data are integrated can we better support business integration and innovation of business models.