An article published by Forbes (sponsored by SunguardAS) details why a large proportion of CMDB implementations fail.

I wanted to complement and provide our perspective on the topic, based on feedback from our market and the capabilities provided by Atrium.

The trends in IT more than ever require a solid control over configurations:

  • Larger, more complex, and dynamic data centers accelerate the risk of bad changes, and push the need for automation
  • Adoption of public and private clouds result in more vendors, more operators, and integration layers
  • The accelerating demand for digital services from the business places IT in tough situations where reactivity and efficiency are key ingredients for success

This drives benefits of Configuration Management beyond what was outlined in the article:

  • Change control/change management: Documenting your environment illustrates the many interdependencies among the various components. The better you understand your existing environment, the better you can foresee the “domino effect” that changing any component of that environment will have on other elements. The end result: increased discipline in your IT change control and change management environment.
  • Disaster recovery: In the event of a disaster, how do you know what to recover if you don’t know what you started with? A production CMDB forms the basis for a recovery CMDB, which is a key element in any business continuity/disaster recovery plan. That comprehensive view of what your environment should look like can help you more quickly regain normal operations.

But also:

  • Automation: With the growing scale of data centers, there is no option but to automate routine tasks. That spans IT Operations Management which need business-driven provisioning, patching or compliance, IT Service Management which need to accelerate incident resolution by efficiently prioritizing/categorizing the work, etc.
  • Performance and availability: With availability being so critical to business success, how can IT be proactive and fulfill SLAs if it cannot map events that impact the infrastructure to the business service that is affected? How can capacity decisions be business driven without an accurate picture of the environment?

The article lists 4 reasons for CMDB failure (competing priorities, limited resources, complacency and overly manual approach).
The fact that a “CMDB Project” is mentioned here is symptomatic that many organizations have initially only considered the technology aspects, rather than establishing Configuration Management as a key discipline that relies on CMDB technology. The human factor is in most cases the #1 source of failure, and there are key questions that cannot be ignored, nor forgotten throughout the implementation:

  • What is the business reason for Configuration Management?
  • What current and future problems is Configuration Management going to address?
  • Who is the sponsor for this implementation?
  • What are the processes that will interface with Configuration Management, either to provide data or to consume data?

This ensures a top-down approach, that starts with a vision, drives the boundaries of the data model, the types of integrations, etc.

Once the implementation has kicked off, there are other reasons that can lead to failure such as:

  • The data getting into the CMDB is not governed correctly: it is Configuration Management’s responsibility to ensure that the data is accurate, and transformed appropriately, so it can be referenced reliably. This needs regular reviews of the rules and filters that automatically govern data accuracy
  • Expecting 100% coverage before going into production is playing on the perception that CMDB fails. Configuration Management is a continuous practice, and CMDB implementations need incremental success because the target will always be moving

When it comes to tips for success, I can’t agree more with the article about the absolute necessity to “automatically update the comprehensive picture of your environment to reflect the potentially tens of thousands of changes per year to your environment.” Atrium Discovery and Dependency Mapping (ADDMDiscovery / ADDM can witness how efficient it is at feeding Atrium CMDB with trustable data, that can be automatically synchronized with service models.

Atrium CMDB definitely provides the most comprehensive solution, in terms of its capabilities to handle incoming data, possible interfaces for data consumers, scalability, and the wealth of integrations that exist with BMC or other vendor products.

Recommended reading: Critical Capabilities for Configuration Management Database (Gartner, June 2014).

A main benefit is that it does not require different tools for different data transformation operations. Now, because of this richness, an implementation has to start with the right understanding of the tool, as well as how it should be used. To that purpose, the documentation includes Best Practices that guide implementation towards meeting success with understanding the data model, loading data, normalizing it and ensuring correct reconciliation with other sources of data.

In a summary, Configuration Management is needed more than ever, and needs to be addressed as a discipline, that leverages the most appropriate tools which will guarantee data accuracy, high levels of automation, and strong integrations to drive the most value.

Atrium is the most widely deployed CMDB so it probably has the largest track record in failed implementations. The other side of the glass also means that it has the largest number of successes. This is confirmed by its users, and the rate of 85% failure is certainly not right when applied to it.