Strategies for justifying CMDB ROI

By Megan Santosus, Features Writer 12 Dec 2007 | SearchDataCenter.com

IT infrastructure news Digg This!     StumbleUpon     Del.icio.us

For more on CMDBs CMDBs: A special report  CMDB software evades data center managers Systems management: Spending tepid, ITIL and CMDB gain credence

• A short-term strategy. Fix immediate problems, such as too many servers, to yield cost savings.
• A midterm strategy. Improve overall service delivery.
• A long-term strategy. Get buy-in for a long-term strategy by arguing that IT cannot best serve the business without being managed optimally; this approach necessitates a CMDB or some similar tool and associated processes.

One of the big benefits of [CMDBs] is return on avoided outages, but that's tough to calculate. Jay Long, manager of IT service management, Forsythe Solutions Group Inc.

The ROI challenges surrounding CMDBs emanate from the very nature of a CMDB itself. In essence, a CMDB is just a repository of data about stuff -- in ITIL parlance, known as configuration items (CIs) -- on your network and in your environment. As a storehouse of information, a CMDB is useful and will therefore will pay off only when the information within one is both accurate and actionable. Consequently, without associated tools such as discovery, inventory and application dependency mapping technologies that enable data centers to monitor their environments in real time or close to it, the value of a CMDB is limited.

"There's a whole ecosystem of technologies that organizations need to leverage other than just configuration tools," said Dennis Drogseth, a vice president and senior analyst at Enterprise Management Associates (EMA) an analyst and consulting firm. In many cases, an organization will already have several of these tools, such as asset management software or inventory and discovery systems. A CMDB can link all these islands of data together and provide organizations with a universal view of a data center environment.

A February 2007 study by Framingham, Mass.-based research firm IDC identified a potential pitfall of CMDB projects: organizations' difficulty in articulating the ROI that justifies the time, money and effort. In particular, the study noted that making the case for a CMDB's ROI tends to focus primarily on short-term cost savings. The problem, of course, is that many organizations will realize savings only in the long term -- that is, once a CMDB has been up and running for a while.

Further such longer-range savings won't become apparent unless you invest the time and effort to develop process standards for change management. Then a CMDB can support improvements in service delivery, which can generate cost-saving benefits. All told, a large enterprise may spend several million dollars and up to three years to get a CMDB fully functional, according to EMA.

• Reducing or eliminating unnecessary servers and other hardware
• Better license and maintenance contract management
• Tighter management of outsourcing agreements
Labor savings

"Many organizations buy maintenance contracts in a piecemeal fashion so they expire at different times," said Tainter. "If you can negotiate coterminous contracts, you can get volume discounts."

At Linklaters, the CMDB was complemented with application dependency mapping software from Tideway Systems that enables the firm to determine relationships among applications and CIs dispersed among four data centers worldwide. According to Gilhooly, the key reason to implement a CMDB and associated infrastructure tools initially revolved around preserving its investment in a disaster recovery plan. "One of the biggest challenges for disaster recovery is making sure that the plan keeps up with changes to the infrastructure," Gilhooly said. "The CMDB can help by highlighting change requests that have a potential impact on the DR [disaster recovery] plan."

As for downtime-related savings, Gilhooly doesn't have precise figures but believes nonetheless that the CMDB more than pays off. "Downtime cost is quite variable for us because lawyers can often work around small outages," Gilhooly said. However, based on some initial calculations, Gilhooly estimated that a complete outage affecting a single data center would cost Linklaters about $715,000 per hour, while downtime for a single lawyer would cost about $408 per hour. "Once the DR plan is broken, the cost of repair would be quite high so it's best not to get into that position," he added.

Initially, Mercy Health used the CMDB to inventory the workstation environment. When the hospital undertook a data center redesign, moving to a new HP BladeSystem server architecture, Giblin opted to expand the CMDB to include server information.

Currently, Mercy Health's CMDB encompasses server information, including information on patch management, inventory, provisioning, software delivery and contract management.

"In terms of server management, our costs are primarily in resource time and manpower," Giblin said. Now information on the hospital's 100 physical servers, such as names of host bus adapters, are included in the CMDB. "We don't have to go out and physically touch the machines anymore," he added. "This takes our server management up a notch."

And Mercy Health hasn't stopped with servers. Giblin said that the hospital now uses its CMDB to track clinical assets --equipment such as respirators, IV pumps and the like.

"These things are nondiscoverable on the network, so we import the information into the CMDB from a third-party database," Giblin said. To facilitate the discovery process, Mercy Health may ultimately install radio frequency identification chips or bar codes on the equipment. Needless to say, this is not how Altiris intended its CMDB to work, but it's a way for Mercy Health to derive more return on its investment.

But the biggest payoff from CMDBs comes when organizations use them to gain visibility into configuration information to make better decisions about change management processes. From a purely operational perspective, Ronni Colville, a vice president in Gartner's IT operations group, said that a view of CIs can result in improved risk assessment and root-cause analysis. In turn, those capabilities can result in more reliable change management processes, and ultimately of IT managing business services.

According to Drogseth, one of the best ways to frame the ROI argument (at least in terms near and dear to a financial analyst) is to look at repair times. "The dominant metric for financial types is reducing the mean-time-to-repair," he said. Drogseth knows of one company that quantified a CMDB in terms of downtime. "The company figured that it cost $1 million per minute of downtime," he said. "The company reduced downtime by 70% just by getting a cohesive view of its environment and reconciling what they had."

According to Drogseth, a 70% reduction in mean time to repair or in downtime overall is not an extravagant goal for most organizations looking to improve IT's operational performance, but getting to that point may require significant time and effort. "It may require months or more than a year of process and organizational planning as well as technology adoption," he said. "Don't forget the core problem here is really the siloed nature of how IT professionals have typically used tools. … This is first and foremost a political problem."

That said, however, Drogseth maintained that CMDBs enable most organizations to achieve significant benefits. "Getting a cohesive, nonredundant and integrated set of trusted sources to support service performance in a disciplined CMDB system can and should produce dramatic value over time," he said.

Still, most organizations don't have the luxury of investing in a CMDB and implementing it purely on the faith that a tool will lead to service management nirvana -- and thereby deliver a solid ROI. IT managers should look carefully at the immediate, short-term and midterm benefits they can deliver along the road to nirvana and determine if there's enough there to build a case for a CMDB project.