05 | D4 model: overview of China Taiwan planning and construction methodology

Last time , I have shared with you four issues that need to be considered before the construction of Zhongtai . Considering these issues clearly will enable us to start the construction of Zhongtai , Avoid some risks in advance .

good , Now let's assume that we have all figured out these problems , How on earth should the middle stage land ？

In the past two years, I have participated in the construction of Zhongtai , I did step on a lot of holes , A lot of detours . Let's use the rest of the second part , Let me introduce you to , At present, we have explored and sorted out the middle platform landing idea in practice , I hope it will be helpful and enlightening to you .

Let's start with , As mentioned in the previous article , At present, the middle stage in the market “ A wide variety ”. Different kinds of middle stage , Their construction methods may be completely different , But there must be some common ideas and methods . In the following part, I will use a Business Center The build process of is a sample , Let me introduce you to the practice of landing in the middle stage , What difficulties will be encountered , Sort out some ideas and methods .

The initial stage of building a typical business platform

In order to let you realize some difficulties and problems in the construction of Zhongtai , We also use the case of geek real estate to simulate a Zhongtai construction from 0 To 1 Start up process of .

After Xiao Wang received the entrustment from his boss , Prepare to start the construction of the middle stage of geek real estate business .

Xiao Wang is a technical and architect , Led the distributed service transformation of several large systems in the company , Very experienced in distributed architecture design and implementation , Domain Driven Modeling, which is often mentioned by Internet companies when talking about Zhongtai , Microservice technology architecture , Also familiar . Yes, of course , This is also the reason why the leaders entrusted this important task to Xiao Wang .

Do as you say , In limine , Xiao Wang is still as usual , Prepare to sort out the business requirements . But the first question puzzled Xiao Wang ： What businesses should be sorted out ？

Before, Xiao Wang dealt with single product level , Just catch the users or business experts of the system , Find out the demand for the product , No matter how complex the product is , There is still a boundary . To be the center stage , It often involves all business lines of the enterprise , Should we sort out all the businesses of the enterprise ？ If you do , Basically, it means mobilizing the resources of the whole company , Why do all business lines cooperate ？ Even if the business will actively cooperate , There are follow-up questions , What granularity does Xiao Wang want to sort out ？

Facing all kinds of problems , Xiao Wang began to feel something was wrong . Technically and architecturally , There is no difference between being a middle platform and being a distributed system , But the situation, scope and complexity , But not at the same level . I don't know where to start for a while .

At this time, Xiao Wang , I always feel something different , But I can't tell what's wrong , Fell into deep anxiety .

What is the difference between being a business center and being a distributed system ？

To be honest with you , This little Wang is me 2017 In, he just started to do the real portrayal of the middle stage planning and landing , Of course, the case is not a virtual case , But the real situation and problems encountered at that time were very similar to those in the case .

And I had a lot more problems than that , For example, even if you sort out all your business , Plan out a middle platform , But how can you prove that you are the right one ？ Is the best ？ It was really a painful experience to recall that time , These questions haunt my mind every day . All the problems can be summed up as a simple problem ： What is the difference between being a business center and being a distributed system ？

Don't bullshit , Or just give the answer . After a long time of thinking and replying , My little friend and I found the key point of the problem , It's very simple , It has been repeated many times before , It's just that the scope is different , If you say it more clearly , Still those three words ： Enterprise class .

Why do I stress these three words over and over again , I know a lot of people don't feel much when they first see these three words , But these three words are really extracted after we stepped on countless pits , Contains answers to many questions .

Now look back , Be a middle stage , In fact, the tools, methods and ideas we used at that time did not have much problems , Just think of the scope and difficulty of the Zhongtai incident as small , Think simple .

Trust a lot of people , Whether it's a mid tier product manager or an architect , Just like I did then , What we do more is system level or single business level system construction or transformation . But when we were in the middle stage , We are dealing with things of a completely higher level and scope , A single product has jumped out 、 A single line of business , It involves the enterprise level .

You may ask , Just a little wider , How much difference will there be , I can only say that there is a big difference .

First , If it is an enterprise level problem , What you need to solve is to achieve the level of enterprise goals . The question itself is vague , Generally, they are at the strategic level , So we can't just start from the current business , Start with the strategic analysis of the enterprise , Fully consider the impact of future architecture planning on strategy .

secondly , If it is an enterprise level problem , What you will face is the organizational problem of the enterprise . Organizational problems are sometimes much more difficult to deal with than business problems , Because it involves the redistribution of enterprise interests . Once the problem of organization and interests arises , There will be all kinds of what I call “ Why series ” The problem of , For example, the most common problem is why we should cooperate with the middle office ？ Why should I give the data to the middle office ？ Why should I use the middle stage ？ And so on .

also , If it is an enterprise level problem , Back to business , It is also completely different from the previous system . What you will face is the overall picture of the business of the enterprise , Even those that will appear in the future , It is not yet known what the potential innovative business looks like .

besides , There will be countless similar difficulties , It is something we have never faced before when making a system and product . That's why I put “ Enterprise class ” These three words have been put into my definition of Zhongtai . Remind yourself all the time , What I am doing is the same as the original system and product , Not a level thing , It's just another species .

Think clearly about the enterprise level , There is also a very, very big inspiration for me , That is, I finally figured out where the wrong places I felt before were ？ actually , I have been using a system level product and architecture approach , Trying to solve an enterprise product and architecture problem . Now look back , It's not surprising to encounter those problems and difficulties .

therefore , The nature of the whole thing has changed , Although we may still do business combing , Be able to do micro services , The same means such as domain modeling will be used , But we need to be clear , When we were in the middle stage , We are essentially building an enterprise level architecture , An enterprise architecture incorporating new elements of the platform , I call it ： User oriented and innovative platform enterprise architecture .

That's the question again , Zhongtai and tradition EA What's the difference ？

After realizing that Zhongtai is essentially an enterprise level architecture problem , In fact, I am not as excited as I thought , On the contrary, I am very disappointed , Why is that ？

Understanding enterprise architecture （Enterprise Architecture） My friends must know , This is not a new concept , image TOGAF This mature EA The framework already has 20 Years of history , We often see, for example, business architecture 、 Application Architecture 、 Technology Architecture 、 The data architecture and even the organizational architecture are contained in EA Within the complete system of .

How can I describe that feeling at that time ？ It's like thinking you've found a new species , It turns out that it's nothing new at all , It's already on record , Printed in every encyclopedia . The sense of loss can be imagined .

not so bad , Later, the story took a new turn . After continuous exploration and practice , I found the traditional EA When dealing with the platform based enterprise architecture such as Zhongtai, the architecture will have a lot of problems , for example ：

• Conventional EA Most of the methods are based on business process , Most of the output is what kind of system to be purchased or developed to solve some problems in the business process , So most of the output is that the enterprise needs to purchase things like ERP、CRM Such a system can solve problems in specific fields . As for how to precipitate into a platform or even a middle platform , It doesn't seem to apply that well .

• Conventional EA The method is more to solve the problems under the information background at that time , That is, based on the status quo （As-is） To sort out our business , Consider how to solve the problem of information transformation of business processes through the construction of the system . At present, when we are building the middle stage , Often the degree of informatization has been very high , All the systems we should have , And the middle office construction is even the digital construction that people often talk about , More for the future （To-be） Business development and innovation , With the traditional EA It doesn't seem that the method of .

• Conventional EA Methods are mostly heavy processes , We need to do a lot of research for a long time , Produce hundreds of pages of planning documents . More than ten years ago, the development of informatization was not high 、 The era when waterfall process is still dominant is also feasible and mainstream , But now the Internet is even a traditional enterprise IT The rate of change , Maybe even if it was planned with great effort , It is out of date , Not really .

therefore , I rekindled my enthusiasm for the research on the methodology of the planning and construction of Zhongtai , I wonder if it can be based on the traditional enterprise framework , Rub in some new ingredients , For example, integrating design thinking （DesignThinking） To solve the problem of innovation , Incorporate Domain Driven Design （Domain-Driven Design） To solve the problem of identifying the capability of the Chinese Taiwanese chemical industry , Then, we can solve the problem of process heaviness by integrating the ideas of agile and lean 、 Long process 、 Low response to change , Gather all the family's leaders , Reconcile a new approach to enterprise architecture , That is, the platform enterprise architecture facing users and innovation .

Twinkling of an eye , More than two years have passed , Up to now, we have groped for an improved version of EA Method , As I said above , It integrates the strengths of each family . At present, we have used this method to help many enterprises to carry out middle office planning , Many others are already in the process of landing . so to speak , This method is very mature , So I'm going to share it with you today .

How is this method ？ We call it “D4 Model ”.

The methodology of planning and construction of Zhongtai ：D4 Model

So it's called D4, The main reason is that we have divided the process from overall planning to implementation and delivery into four different stages , Including two divergence and convergence processes .

The first stage is Discovery, Help us build a global vision before planning for the middle office . In this process, we take the enterprise vision and strategy as input , Combined with industry trend 、 Competitor analysis 、 Customer group analysis 、 Business status analysis 、IT Asset inventory , Understand the strategic market environment, business and IT Full picture , Help us make the most correct judgment .

The second stage is Define, Help us build on the previous Discovery Convergence and analysis of divergent dimensional information , Define the architecture of the midrange . Analyze the coincidence degree by combing the business across business lines , Combined with the domain analysis, the core problem domain of the enterprise after the business representation is further expanded and the coincidence degree is analyzed , Let's converge and deduce the enterprise architecture design based on the middle office . And score based on multiple dimensions , Form specific implementation path planning , To put it bluntly, what you do first and then what you do . Here's a little bit of attention , At this point, the convergence is still at the enterprise architecture level , Like the middle office in business 、 Products of this level in the data center , It may just be a project in the implementation path , At this stage, we can also answer the question we care about , Do we need Zhongtai or not , What kind of middle stage do you need ？

The third stage is Design, Help us target a product in the implementation path , For example, business center , Do detailed design , Including product level business requirements analysis 、 Function and architecture design 、 Implementation plan, etc . For example, for products in the business , stay Design At this stage, we need to answer the product vision 、 The border 、 Product form 、 Technology Architecture 、 Delivery plan 、 Cost estimate, etc , This process is a standard product design process , However, most of the mid-range projects are aimed at mid-range products .

The fourth stage is Delivery, At this time, we can aim at a designed middle platform , Start the specific delivery process , We use agile and lean software development , With fast iterations and feedback based tuning , Make up for the design deviation and other delivery problems caused by the complexity of Zhongtai construction to the greatest extent , And pay attention to the governance and protection of the architecture , Reduce the deviation between implementation and design .

Whole D4 The process , It is a strategy from strategy to implementation , The process from enterprise architecture to final delivery of product architecture . And follow the ideas of agile and lean , Whole D4 The process of is also iterative , For example, every quarter to half a year , We can do a light weight again Discovery and Define, To continuously make agile adjustments to the enterprise architecture , To cope with the changes and uncertainties of the market and itself .

Last , A classmate asked me , Why call D4, Don't cry 4D？ In fact, there is a small colored egg here , Because we think D4 The pronunciation in Chinese is a bit like Diss, It represents our constant challenge to the old business form 、 Continuous innovation 、 A changing attitude , It is also very consistent with the current wave of enterprise digital transformation .

To sum up

Today's talk about , Let me take you through a typical problem that will be encountered in the start-up stage of the middle office construction , The difference between being a middle platform and being a distributed system is described in detail , And finally understand the essential problems behind Zhongtai , It's actually A user oriented and innovative platform enterprise architecture .

after , We also analyze why the traditional enterprise architecture method has limitations in dealing with the middle ground problem , Combined with other relatively new practices , It is integrated into our current planning and construction methodology for the middle office ,D4 Model .

Last , I gave you a brief introduction D4 The whole picture and idea of the model . So far, we have passed this methodology , In helping a number of enterprises to carry out the planning and implementation of the middle office .

From the current feedback effect , This model is very easy to use , Because this method actually implements what we call Think Big,Start Small,Move Fast Principles , Think long-term , And cut in quickly , And keep evolving . When dealing with the planning and construction tasks of the middle office with extremely high uncertainty and complexity , Such a principle is particularly appropriate and necessary .

Last , I'll leave you some questions to think about ：

• In the process of the construction of Zhongtai , What kind of problems have you encountered ？ How to deal with it ？

• Your business , What kind of methods are being used to plan and build the Zhongtai ？ And what I said D4 What are the similarities and differences between the models ？