Skip navigation

Overhaul Precedence Engine for Overlapping Data

score 110
You have not voted. Product Team Review

I believe it is time to overhaul of the precedence engine to support todays’ customer needs.  This isn’t a small enhancement
but a product design consideration.  The features of the Atrium CMDB precedence engine make a significant assumption that customers don’t have multiple sources of data that overlap.  This works fine when you have one source of data for particular CIs and you accept that the data source is not perfect and you are willing to live with less than complete or accurate data (i.e. you live and die by you single source of data).


This worked for us several years ago, but in today’s world the CMDB is being usedfor IT Asset Management (ITAM) and Software Asset Management (SAM) and data quality is paramount.  Within the ITAMand SAM processes incomplete or inaccurate data can have significant impact to an organization financially and procedurally. Giving management the excuse we can support one source of truth is no longer acceptable.


The precedence engine should evolve to support overlapping sources of data.  I believe the time for this is now, indicated by a new
segment of products (e.g. Blazent’s Configuration Manager) are being marketed to replace/augment the Atrium CMDB precedence model.


A few requirements I would like to see possible in the precedence engine are:


  1. Ability to select the best data source for specific data elements for a sub-section of data using a qualification. 
    (e.g. Today the precedence rules are at the class level.  We need to be able to set precedencefor CIs in the computer system class which have a Product Categorization of X,Y,Z to use SCCM data for the environment, host name, and domain field instead of
    the ADDM value.  Another example, VMWare Virtual Machines need to use the values from Bladelogic in a rather than ADDM or SCCM)
  2. Ability to rank sources by class or element to derive the best choice.
    (e.g. If two sources agree and one does not use the value the three agree on.  ADDM, SCCM, and Bladelogic are all
    providing Operating System data.  If ADDM is the primary source of data but Bladelogic and SCCM share a different value
    make the SCCM/Bladelogic value win)


I would appreciate any comments on this post from other end users who have similar challenges and might be able to contribute to the list of requirements.


Vote history