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

[Rodel]

Hey everyone,

I wanted to check and see if anyone has successfully implemented the CMDB without using ServiceNow Discovery and Service Mapping and what did it look like?

• Do you manually create Business Services and Business applications?
• Do you automate the creation of some assets using the various data sources?
• Do you map the applications to the server it resides on and the network device it uses manually? 

I am waiting on budget to approve the purchase of Discovery and Service Mapping, but I want to know my options if this is something that will not be approved? What is the minimal viable product that works for your organization? 

[Rodel]

Hey Dave,
Thank you for sharing the CSDM 2.0 white paper. I started to mock up some hypothetical services for “Communications and Collaboration Services” and had some questions. I am working on the best approach to implement the “Crawl” method in the white paper that focuses primarily on Business Applications, Application Service and Application (Discoverable) and Server/Host (Discoverable). We plan on using SCCM to support the discovery of Server and Hosts without the relationship’s day 1.

[Rodel]

I attached my message and graphic so that it is a little easier to read. 

[David104]

I agree that some level of discovery is essential to the success of a CMDB. You can build processes to capture CI's if there is a central channel for provisioning, but inevitably someone will provision outside of the process and the holes in your data will grow.

From a service mapping point of view, I've had limited success in the past using provisioning and decommissioning processes to capture relationships, but again it is only valid if those processes are used consistently.

Whether you are using Discovery and Service Mapping, or relying on process to capture CI's, you should always ensure that the CI Class owners are well defined and that they regularly review the CI's that they own to validate the lists and ensure that the correct relationships are captured. By distributing the responsibility, the workload for CMDB Admins is reduced and the owners can no longer complain that the data is not accurate because it's their own fault. The CMDB Dashboards are great for keeping an eye on that too 🙂

So as per the above post, it's possible but so much more effort is required for a lower quality result.

[DaveHertel]

Agree 1000%.   And your points about distributing the load of ownership (CI / classes / services, etc.) is spot on -- by putting ownership and responsibility in the hands of the people who have the strongest interest, this keeps data quality higher and therefore usefulness of the CMDB increases.   Automation via Discovery is by far the best way to keep the core infrastructure data updated IMHO 🙂

[Damhoej]

Yes it is absolutely doable 🙂
I am building a CMDB basically from scratch without ServiceNow Discovery.
I do use MDR (Main Data Repositories) that I import into ServiceNow with a scheduled import.

If you are on Orlando then this feature looke really cool: https://youtu.be/YvsCY0M7JLw

You have to get into CI Class manager and start exploring all the classes and the subclasses that you need.
There multiple paths to take, but start simple..
Start from asset management and get your hardware created correctly (through Asset Management and Procurement). When you got the hardware done, then go on to virtual servers.

Underneath I have drawn an example of the CMDB for a Hyper-V topology.
Each box should have its own CI in the CMDB, however for assets you should create them first.

If you would like to know more about assets and models (very important) I can recommend these articles:

https://community.servicenow.com/community?id=community_blog&sys_id=4fdd2ae9dbd0dbc01dcaf3231f9619ef