Is the division of each capability domain of Dama, dcmm and other data management frameworks reasonable? Is there internal logic?

At present, the data management system framework is relatively complete, including those of the International Data Management Association 《DAMA Data management knowledge system guide 》、 Ministry of industry and information technology 《DCMM Data management capability maturity assessment model 》、 It's in the hospital 《 White paper on data asset management practices 》 etc. , For the convenience of the following description , Hereinafter referred to as 《DAMA》、《DCMM》 And 《 white paper 》.

《DAMA》 This paper gives its own framework for the data management system , Here's the picture ：

《DAMA》 To determine the 10 Data management functions , That is to say 10 Three capability domains , Namely Data architecture management 、 Data development 、 Data operation management 、 Data security management 、 Reference data and master data management 、 Data warehouse and business intelligence management 、 Document and content management 、 Metadata management 、 Data quality management And Data governance .

Novices must be confused when they see this framework , On the one hand, this 10 There seems to be no logical relationship between the two functions , Not at all , On the other hand , Data governance as a separate function , I wonder what the difference is with management , Even though we sometimes bite the bullet , But once the very similar words of governance and management are put into the same context , It will confuse most people .

that , Look again. 《DCMM》.

《DCMM》 be relative to 《DAMA》 Easier to understand , Including 8 Three capability domains 29 Ability items , Namely Data strategy 、 Data governance 、 The data architecture 、 Data applications 、 Data security 、 Data quality 、 Data standards 、 Data lifecycle , As shown below ：

And 《DAMA》 comparison , There are mainly the following differences ：

1、 Data strategy is separated from data governance , Become a capability domain alone , This is a matter of different opinions .

2、 The data application capability field is added , It should include DAMA Business intelligence management , But it has a broader connotation , It also includes developing shared services and so on , This is mainly influenced by the background of the times , After all DAMA When it was released , The application value of open data sharing is far from being recognized and understood .

3、 The data standard capability field is added ,DAMA Reference data and master data management in are included in its scope , That is to say DAMA The master data and reference data of are degraded , But this is a matter of different opinions .

4、 The data lifecycle management capability field is added , This is a good generalization , It's actually a DAMA Data development in 、 Data operation management 、 Data warehouse management is included , But more comprehensive , And more logical .

5、 Data architecture adds new meaning , It includes metadata management capability items , be relative to DAMA, Metadata management has been downgraded .

Last , Look again. 《 white paper 》.

《 white paper 》 Include 8 Management functions and 5 A safeguard ,8 Management functions include Data standards management 、 Data model management 、 Metadata management 、 Master data management 、 Data quality management 、 Data security management 、 Data value management and Data sharing management ,5 The three safeguards are strategic planning 、 Organizational structure 、 Institutional system 、 Audit system And Training, publicity and implementation , As shown below ：

And 《 white paper 》 comparison , There are mainly the following differences ：

1、 Metadata management and master data management are separated from data standard management , Become an independent capability domain

2、 Data architecture management is gone , Only one data model management is retained

3、 Data application management is gone , Only one data sharing management is left

4、 More data value management , It is mainly used for data value evaluation

5、 Data governance is gone , Replaced by safeguards , Including strategic planning 、 Organizational structure 、 Institutional system 、 Audit system 、 Training, publicity and implementation

Through the above comparison, it will be found that ,《DAMA》、《DCMM》 And 《 white paper 》 Although they are all talking about the data management framework , They have a lot in common , But there are still many differences , that , Which one should be adopted to best reflect the connotation of data management and make the logic more self consistent ？

Personally think that , The data management framework should consist of two parts , The first category is Data management activities , The second type is Data governance , That is, the control activities to ensure the normal development of data management activities , The so-called management of management ,《 white paper 》 I'm afraid we don't understand data governance , Data governance has also been changed into a safeguard measure .

But whether it is data management activities or data governance , We all want an answer , namely Is the division of these capability domains reasonable ？ Is there a logical relationship between these capability domains ？

Let's start with data governance .

《DAMA》 Data governance for includes strategies 、 organization 、 policy 、 Review and other contents ,《DCMM》 Data governance for includes organization 、 The system 、 Communication, etc , Plus a separate data strategy ,《 white paper 》 Our safeguard measures include strategies 、 organization 、 System, etc , You can see it , The three are basically consistent , And the logic is relatively simple .

《DAMA》 The data governance content described is actually the most comprehensive , But a little verbose , Not easy to understand and record , I have the courage here , Divide the core functions of data governance into strategic planning 、 Organizational security 、 Policy formulation, review and communication , But culture is not written .

Strategic planning ensures the right direction of data management activities , Organizational assurance ensures that data management is performed .

Policy formulation ensures the principles of data management activities , Review communication to ensure relevant specifications for data management activities 、 Standards and processes are empowered and implemented by the organization .

Let's talk about data management activities .

from 《DAMA》、《DCMM》 And 《 white paper 》 You can see , The capability areas of data management activities are not exactly the same , Everyone has their own focus , Probably because all kinds of data management activities are produced to solve actual data problems in a specific context , Not the result of abstract summary , As a result, the boundaries between various data management activities are not very clear , There may even be a crossover .

such as 《DAMA》 Including master data and reference data management , It does not include data standard management ,《DCMM》 Including data standard management , Master data and reference data exist only as part of standard management , There is obviously a logical problem , Because the connotation of master data and reference data management is far beyond the data standard ,《 white paper 》 Made a compromise , Manage master data 、 Reference data management and data standard management are juxtaposed , But obviously the content is overlapping , Because master data management 、 Reference data management must also include standard management .

DAMA Data standard management is not listed as an independent data management activity , Maybe it's because it's not rigorous , But why 《DCMM》 And 《 white paper 》 Or will data standard management be listed as an independent data management activity ？

It may be related to the importance attached by the domestic financial industry to data standards , Because the financial industry is strongly regulated , Data standards are needed as the starting point for supervision .

Huawei's way of data is talking about the base of the data lake , The following architecture diagram is given , Data asset management on the right lists many data management activities , You will find that indicator management is juxtaposed with metadata management .

Asked why , Only then did I know that one of the business pain points of Huawei's data management was the indicator problem , Therefore, it is specially marked in the architecture diagram , Not rigorous but very practical .

The same problem occurs in metadata management ,《DAMA》 and 《 white paper 》 Both regard metadata management as an independent capability domain , and 《DCMM》 But put metadata management under the data architecture , This logic doesn't make much sense , Because metadata management is related to data architecture , But it is certainly not the relationship between inclusion and inclusion .

Metadata is the overall management of data description information , Not just a description of the data architecture , It also includes the description of all objects such as data quality , Why? 《DCMM》 To do so ？ Maybe with 《DCMM》 Related to its own use , After all, it is used to assess the maturity of data management , Need to be able to easily assess , Perform in place .

《DAMA》 The division of capability domains is rigorous , But we can't surpass the times . For example, at the data application level ,《DAMA》 Only business intelligence is mentioned , But in the age of big data , The word business intelligence can no longer cover all the connotations of applications in this era 《DCMM》、《 white paper 》 Mentioned the data application , Data sharing, etc , This is the result of advancing with the times .

Above all , Here I give a preliminary framework for new data management activities , Include 8 Management activities ： Data quality management 、 Data architecture management 、 Metadata management 、 Master data and reference data management 、 Data security management 、 Data lifecycle management and data application management .

that , this 8 Are the data management activities logical to each other ？ Why is this 8 A? ？

The value creation process of data is divided into data generation 、 Data processing and data consumption , The key data management activity corresponding to these three stages is data architecture management 、 Data lifecycle management and data application management .

Data architecture management ensures that data is properly designed and generated .

Data lifecycle management ensures efficient data collection 、 Storage 、 demand 、 Development 、 O & M and destruction .

Data application management ensures that data can be fully shared 、 Open and service .

In order to ensure the smooth progress of these three stages , There are four basic management activities that go through each stage , Data quality management 、 Metadata management （ Data standards management ）、 Data security management 、 Master data and reference data management .

Data quality management 、 There should be no objection that data security management should run through the value creation process of data .

Master data and reference data management are the same , For example, when designing the data architecture, we should fully consider , Application construction is more inseparable from master data and reference data .

I am a little hesitant about whether metadata management should be included , Here's why ：

First, the concept of metadata is too big ,bit The world except the data itself , The rest is a description of the data , That is, metadata , All activities of data management have metadata management content .

Second, metadata is a pure technical term , It is easy to lose focus if it is too general , You talk to others about metadata management , Others simply don't know what problem to solve ,DCMM Simply put metadata management under data standard management , Emphasize the attribute of standardization , therefore , It is also reasonable to replace metadata management with data standard management , Although metadata management has other contents besides data standard management , But the content is too scattered , It is better to apply it on demand in activities such as data life cycle .

If metadata management should emphasize the unified management of metadata , I don't think so , Nor can it be unified at all , This has been proved by practice . For example, many so-called metadata management systems , Only the data directory is included 、 Consanguinity analysis 、 Limited functional modules such as impact analysis , But is it cmdb Not metadata ？ I haven't seen you either olap and oltp The metadata of has been managed uniformly .

that , Why are these four basic data management activities , Is there anything else ？

I can't answer that , Basic data management activities should be abstracted from common problems encountered in practice , Perhaps with the expansion of data application , And new activities will be summarized , For example, only when data elements flow can they play a greater role , Perhaps in the future, open data sharing can become a data management activity alone .

Based on the above analysis , I use the following data management framework diagram to make a summary , Data management “434” New framework .

"4" The four aspects of data governance , It can ensure the smooth development of data management activities .

“3” Namely, the three data management activities involved in the process of data value creation .

the last one “4” Namely, the four basic data management activities throughout the process of data value creation .

I hope you have a clearer understanding of the internal logic of data management .