Idea ID: 1686337

Split Computer CIs into Server and Desktop

Status : Under Consideration
Under Consideration
See status update history
over 2 years ago

It is quite often I get asked, how many servers do we have, how many desktops. It woud be nice if the Desktops and Servers were not lumped together in one CI. The Node is Desktop attribute is not always reliable enough to distinguish between the two. This might need to be split further beyond just Desktop and Server though if there are other iterations that need to be included. 


  • Are they really breaking it up on a data model level? Honest question, because I don't know it.

    But having a look at our ITSM tool (not platform): SMAX also breaks it up. At least in the SACM module I can see Servers and Workstations. But it seems, that this is also done just on a visual level, because I can also see all "Devices", which includes everything.

    From a ITSM platform perspective: The UCMDB of course is part of that ITSM platform, agreed. But it's not the typical end-user tool from my perspective and also a new user should get a little introduction that explains the TQL concept (which is essential for the whole tool) and therefore he should not take more than 1 minute to build that TQL.

    Which brings me to another idea: If that report is a report, which is needed very often by different user, then the UCMDB administrator could just build it and provide it to all users.
    Or - same idea but different approach: As the end-user is intended to use the CMS UI / Browser, Micro Focus could just implement this as a basic report, which is available ootb, so the end-user does not have to make any configurations and just hit the button.

    However, changing the overall data model underneath and making it more complex is not a good idea I think.

    But despite the way it will be implemented: The real problem is somewhere else - the attribute value will have to be set correctly.

  • If you look at other leading ITSM platforms, they all breakout Computers vs. Servers.  If a new user wants to generate a report of desktops, how long would it take them to build a TQL filtering on the relevant attribute vs clicking on "computers" or "desktops" and exporting?

    I understand there are multiple ways to accomplish this with TQL's and filtering, but I believe breaking out these CI types makes this system more intuitive.

  • Stumbled across that question during a recent customer workshop (and now here).

    I copy Daniel, Glen and Ivan: The possibilities are already covered by the data model and the possibilities of the UCMDB. It's just a matter of data and/or discovery quality if it is usable. But this will be just the same, when adding new CITs to the data model, so the goal should be rather: improve the quality of data getting into the UCMDB to make it usable (building a TQL filtering on the relevant attribute is no rocket science afterwards). 

  • fully agree with Ivan and Glen.

    if the "Node is Desktop" attribute is not reliable in some situations, it will not help to just add a new CIType. instead the discovery has to be fixed.

  • I agree with Ivan, this is a bad idea.

    The complications of adding new sub CIT of Windows or even Unix, in regards to identification and reconsiliation.

    If you have the same CI's comming in from a multitude of sources, you would have a CI Type Flip/Flop scenario.

    Use node_role as suggested, and spent the energy on improving the detection of those values.

    Its free for all customers to create own CIT, if its fits there scenario.