Is it possible to implement CSDM without Discovery?

[Thiagofmeira]

Hi All,

I have a customer and they are very worried about how to structure the CSDM and are rushing to create the services etc.
But they still don't have many things in place, for example the Discovery or the CIs. At the moment they want to import CIs manually.

I my point of view, they are spending too much energy in something that should be analyzed in the future, when we have the foundation data in place and at least a Discovery running in their network.

I would like to hear your opinion about that and understand if am I wrong.

Thank you!

[Aman Kumar S]

Hi @Thiagofmeira ,

Its beautifully explained in below article:

Can you have a successful CMDB without Discovery and Service Mapping? 

[EricDohr]

Hi @Thiagofmeira 

I am assuming the value of Discovery is clear to you, but maybe not all the personas involved.  That would be step one for me to not focus on how it works, but the value proposition it can provide.  

I would recommend taking a step back and answering the following key questions

• What are the business imperatives that are trying to be addressed
• What products are in scope (ITSM, ITOM, etc)
• What are the personas that are in scope

Based on this, you should help you drive your CSDM journey and guidance of the importance of Discovery.   Map out the various value of Discovery in place and see how that fits against your requirements.  Depending on the organization size, number of CIs, team doing the data governance, etc., this will vary from company to company. 

For example, if the business imperative is to "Minimize business-facing services by managing critical incidents by proactive monitoring,” discovery is going to be imperative for the Monitoring team, Service Owners such as the Server team, and products ITOM and ITSM.  If the business imperative is “Provide a quality employee experience by providing self-help through knowledge management,” discovery would not be as influential for the Knowledge Manager, Service Owners, and ITSM product. 

Remember, CSDM is a journey and not a milestone.  What is the cost of not having discovery, maintaining manual records/data quality, and impact to processes will help shape potential risk impact. 

Best of luck!

[SN_Marcel Horn]

The simple answer is yes.

Having a solid structure of application and services defined makes it a lot easier to map your CI to it.

Because in general you should start with a vertical discovery first while identifying your Business Application & Business Services / Service offerings. You going through the conceptual- and business layer and hitting the ground of your CI Level.

Your discover journey could be run in parallel.

@EricDohr  bring up the focus topic, which is important as well.

(What does your company / your architect is trying to accomplish first?)

Hope this helps.

Best
Marcel

[CMDB Whisperer]

There are a few different ways to answer, depending on the trajectory of the CMDB implementation.  

It is possible to define a CMDB that is primarily concentrated on the conceptual layer, defining the application services without the infrastructure dependencies, which will still enable some capabilities in managing services even without the infrastructure CIs.  So if the purpose, for example, is to gain an understanding of the services associated with incidents and changes, then you can still do that without knowing the specific servers involved, although you will of course not be able to manage those dependencies or understand the impact of infrastructure changes and outages without those CIs.  

However, if like many or most companies you wish to understand the use of your physical infrastructure and the dependencies on delivering technical and business services, then your focus should definitely be on integrating your technical data sources and/or implementing Discovery to obtain trustworthy data about your infrastructure CIs.  Although the infrastructure itself is considered part of Manage Technical Services domain in CSDM, it is arguably closer to Foundation in that sense.

So I agree in the context that you have provided that focusing on CSDM Crawl while populating your Application Services by manually importing CIs is problematic.  While manually relating the CIs to Application Services can be a workable (if tedious) approach to populating services, manually importing the CIs themselves is unreliable at best, and will undermine the value you would obtain from your CMDB.  I would strongly recommend focusing on horizontal discovery prior to building the service later.