IT managers who want to resolve problems more quickly, manage the lifecycle of software and hardware assets better, and adopt IT governance best practices have probably thought by now about implementing a configuration management database.
CMDB has become the buzzword du jour for management software vendors and ITIL advocates. CMDBs promise financial gains, through better asset management, and reduced manual labour, because of automated processes, to increase operational efficiencies derived from management data correlation across multiple systems.
However, what is actually involved in deploying an effective CMDB eludes many potential customers, experts say.
“There is a lot of confusion over what a CMDB actually is, but it is such a compelling proposition that IT managers must wade through that confusion to find their way towards the management data integration and best practices it will provide,” says Dennis Drogseth, a vice president with analyst firm Enterprise Management Associates.
Born from ITIL best practices, a CMDB includes a large amount of information about all IT configuration items within a corporate network. That data is drawn from various sources, including applications, operating systems, patches, hardware models, life-cycle costs and end-user connections.
CMDB supporters say the technology helps IT managers aggregate data from multiple sources into a single system and make sense of different data formats, and reduces the time it takes to solve problems. For example, if a server is downed by a configuration change, the CMDB could, in theory, identify the change quickly and help IT managers restore the system before there’s an impact on service levels.
Today, most companies’ configuration data is stored in different formats throughout the organisation, on desktop machines, servers, patches, operating systems and network devices. A CMDB would consolidate such data into a single repository and break down the walls between existing IT domains, such as servers, storage and networks.
However, because the idea of collecting and storing configuration information about every IT asset into one database is overwhelming — and, for the most part, impossible — industry watchers advise IT managers to look into linking multiple databases together into what’s known as a federated database. A federated database lets customers collect data from multiple sources without having to store it all in a single, monolithic database. Federated databases let data reside in multiple sources, with a centralised source knowing where the data lies throughout the company.
“In its simplest form, a federated CMDB is where all the information about your assets lives. The data itself may not be there, but there will be a record of where the data associated with each asset is, what has changed and when, and where to go to find it in your network,” says Jasmine Noel, a principal analyst with Ptak, Noel & Associates.
“People realise they are having availability and performance issues, and the reason is [that] someone changed something. A CMDB, ideally, would quickly point out what has changed.”
Management software heavyweights BMC, CA, HP and IBM have delivered products designed to support customer CMDB initiatives. Other companies, such as Managed Objects, Opsware and Tideway Systems, have equipped their management software applications to feed data to a CMDB. Although they’re enabled by open interfaces among management systems and federated database tools, CMDBs are less about technology and more about process.
“It is a multidimensional resource for network staff, application developers and systems administrators to do fault, performance, change, asset and service management,” EMA’s Drogseth says.
“A CMDB underpins many IT initiatives and, in the real world, is more about integrating management data to improve operations than it is about best practices.”
Early adopters of the technology confirm its potential benefits and challenges.
Simon Gilhooly, head of global technical systems at Linklaters law firm in London, is implementing a CMDB to support IT inventory, disaster recovery, hardware negotiations and service management initiatives. Gilhooly says products such as HP OpenView, BMC Remedy and software from Tideway Systems will contribute to the CMDB, which is part of the company’s ITIL adoption efforts. When the CMDB is established, the vendors’ products will feed data to it, or the CMDB will know to search those product repositories for it.
“The CMDB will need multiple sources of information from our hardware, services and systems, as well as products that collect configuration and dependency information across our datacentre, like Tideway,” he says.
“We want to be able to classify and create a management dashboard of major applications by region, and their availability and performance history, for example. The CMDB would be part of that.”
For Lucent Technologies, a CMDB became part of a larger IT service management initiative the company embarked on with HP last year. Sheila Bridge, who works in IT operations control at Lucent, says the joint project let the company reduce help desk costs by 50% and the time it takes to close tickets by two-thirds.
“We knew it was going to be a challenging project because we were trying to do a lot, including adopting 12 processes,” Bridge says. “We are now in a position to quickly determine the relationship between customers, tickets and IT assets within our CMDB.”
When faced with retiring a service desk product and purchasing a new one, Frances Findley kept her organisation’s plans for a CMDB in mind. Findley, project management analyst for information services at healthcare provider MultiCare Health System, says CA’s support for the configuration management initiative helped her decide on CA’s Unicenter ServicePlus Service Desk, which includes the company’s CMDB. The CA CMDB is composed of a central, federated database with hooks — CA calls them Universal Federation Adapters — into other data sources.
It does not require IT managers to abandon existing databases and move configuration data to another server; it lets data reside in multiple places throughout the enterprise, with the central database knowing where that data is.
“We now have a CMDB with the ability to see trends, [with] defined metrics, automatic notification and issue escalation,” she says.
“It has allowed our department to look more at the overall process and not specific individuals.”