In-Depth

Maximizing ITIL Standards: Choosing a CMDB Approach

With so much industry buzz around configuration management databases, why are so few actually in production?

• By Pierre Coyne
• 10/25/2005

Most enterprises have heard of the ITIL by now. The Information Technology Infrastructure Library’s set of best practices promises to bring unprecedented levels of efficiency and effectiveness to increasingly complex IT environments. In fact, the two most important goals of the ITIL standards are to help IT managers better document IT actions for audits, and to better align IT to specific business objectives.

To achieve these goals, enterprises are putting the ITIL’s best practice recommendations into practice often starting with the incident, problem, configuration, change, and service-level management processes. Undoubtedly, compliance regulations such as Sarbanes-Oxley have been a key driver of ITIL adoption. Government regulations are mandating independent analysis of operational risk by auditors and imposing the threat of financial penalties—they even call for jail time for executives whose companies fail to comply. For CEOs, CFOs, and other executive management, the need to understand the link between IT and the business has never been more crucial.

The flurry of mergers and acquisitions across many enterprise segments has also accelerated ITIL adoption, as organizations look for ways to manage down operations costs, and improve competitive advantage and customer retention through higher service quality.

CMDBs: What You Don’t Know Can Hurt You

As enterprises consider or begin to implement the ITIL’s guidelines, how they approach a Configuration Management Database (CMDB)—the heart of the ITIL's standards—must be strongly considered. As a result, there has been increased discussion on CMDBs over the past year, significantly elevating its stature and priority with enterprise CIOs.

According to Gartner, “As organizations recognize the importance of controlling and tracking changes within their environment and search for solutions, they are looking for a central database to maintain a consolidated view for every element, including relationships (which are often peer to peer) and dependencies (which are hierarchical) within and across elements that make up an IT service.”

With so much industry buzz around the CMDB, and so many vendors touting its virtues, why are so few CMDBs actually in production?

There are numerous challenges—some technical and some political. As defined by ITIL, a CMDB contains all relevant details of each configuration item (CI) and details of the important relationships between CIs. Further, the CMDB is likely to be based on database technology that provides flexible and powerful interrogation facilities.

Several schools of thought exist about the optimal way of implementing and architecting a CMDB. In essence, there are three fundamental approaches.

• In the centralized CMDB approach, all CI details are stored and managed in a single, central database.
• With the federated CMDB approach, data initially resides in individual point tools and is imported into a central database.
• In virtual CMDBs, data is accessed from individual point tools (as in the federated approach). However, unlike the other two approaches, data in a virtual CMDB is not imported into a central database; rather, it is dynamically accessed as it is needed.

Knowing What’s Best for Your Organization

What are the primary roadblocks to CMDB success? What is the right CMDB approach for your organization?

In most enterprises, CI data is stored in a mix of discovery, inventory, and other specialized tools, owned and managed by individual silos. By design, a centralized CMDB and a federated CMDB require data to be stored in a central database. In the first case, a slash and burn approach is required, raising political flags as well as technical challenges; data owners must facilitate a complex and often expensive data-mapping and migration exercise, give up their individual tools and data stores, and adopt a new set of operational procedures.

As Gartner observes, “Because configuration management touches so many IT functional areas, it can be a politically difficult challenge.” It also notes, “Through 2009, for large companies, a single companywide configuration management database solution will be unattainable.”

The second approach (using a federated CMDB) promises to ease some of these political issues in that individual silos can retain some of their point tools. In practice, however, many of the large vendors (including BMC, HP, and CA) that are proponents of the federated CMDB approach currently offer limited integration with third-party tools, instead promoting their own set of point product offerings as an alternative.

Many enterprises that have gone down the federated CMDB approach, whether they chose a large vendor’s federated CMDB offering or developed their own in house CMDB solution, have realized that the challenge of integrating all of their disparate data sources, each with different data schemas into a single common CMDB schema, is cost prohibitive in terms of consulting dollars required and time-to-value.

Forrester has reported that it “strongly believes that the term ‘a CMDB’ is a complete misnomer.” The firm adds that, “No vendor is capable of developing a single database that holds all the relevant information in the required format at the same time and that can scale to the needs of larger corporations.”

The latest arrival—and perhaps the most innovative of the three approaches—bypasses many of these common political and technical challenges. Rather than focus on delivering a central repository where all relevant information is stored, the virtual CMDB focuses on another key aspect highlighted in ITIL’s definition of the CMDB. It aims to provide “flexible and powerful interrogation facilities” as the foundation for visualization and analysis of distributed CMDB data.

In doing so, the virtual CMDB need only normalize data from distributed data sources for presentation, rather than undertake complex data migration or costly common schema initiatives. With data residing in their native data stores, scalability issues associated with importing vast amounts of data into a single database are all but eliminated, as are the political challenges associated with tools replacement, training costs, and complete operational process reengineering.

Picking an Approach

There is no question that ITIL’s CMDB provides the foundation for visibility and actionable intelligence needed to streamline IT operations and accelerate business in the enterprise. The real question is how long will it take for enterprises to realize these benefits, and at what cost.

Enterprises will need to define their objectives carefully, and evaluate the CMDB tools that are most appropriate for their environment. Small organizations can consider a central CMDB approach, as it may provide the right subset of information they need to manage services more effectively.

For larger enterprises, the federated CMDB approach can provide some of the structure and process control enterprises require in the long term. The approach has a heavy startup cost (in terms of time and dollars invested), and there are potentially insurmountable roadblocks (such as scalability and political issues).

Alternatively, or in tandem, medium and large enterprises should consider a virtual CMDB approach as a means to deliver the visibility, workflow efficiency and cost savings they are looking for from a CMDB initiative in the short-term.