You may be installing virtualisation tools so that one server can do the job of five. You might be using configuration management tools to swap applications from one machine to another, depending on load. You may simply be looking to retire old hardware in order to run apps on new, more energy-efficient multi-core systems. But how do you even know what's out there to consolidate?

In large, dispersed environments, identifying consolidation opportunities can be a time-consuming job, requiring the combined efforts of engineers and systems architects working with everything from asset management tools to network discovery applications, to performance monitoring utilities, to home-grown spreadsheets, to big, old-fashioned whiteboards in order to determine what pieces of your hardware and software infrastructure might be better off someplace else.

But a new segment of products -- sometimes called data centre intelligence or consolidation management tools -- are promising to help you automate the process, freeing up the time of some of your most valuable employees while simultaneously providing the hard numbers you may need to justify a consolidation project.

And while these tools now come primarily from smaller vendors, the big guys are gearing up to include the functions in their own business systems management suites.

Bell Mobility, one of the largest mobile phone service providers in Canada, had a problem. More accurately, it had one problem (or crash) after another. Recovering some critical systems could take hours, disrupting services and costing the company serious money -- $2.1 million per day for one system alone.

So in 2005 Bell Mobility launched a study to find a way to speed disaster recovery. One of the recommendations that emerged was to consolidate applications and servers in order to centralise recovery efforts, with the hoped-for side effect of improving server utilisation.

At the beginning of the study, Bell brought in a consultancy to evaluate its operations and offer suggestions for where consolidation made sense. Michel Tremblay, manager for OSS network engineering at Bell Mobility, thinks that's a good way to go. "If you support those systems, it's hard to say, I'm going to cut off my system by so much percent," he says.

It's more likely, according to Tremblay, that an internal IT staff with relationships to the systems might be tempted to say, It's working, so why should I do this?

Disaster recovery benefits aside, estimates showed that if the company could consolidate its server pool by 25 percent (its goal for 2006), it would save $1.5 million in hardware replacement expenses over the next two years -- even without including ongoing support costs for the systems that would no longer exist.

Numbers like those put Bell on a path to find a tool that could help it identify which systems were ripe for consolidation. Capacity analysis and asset management tools could do part of the job, but Bell found a product from a small vendor that seemed to address the heart of the issue.

Tool talk

Bell originally had used startup Cirba's Data Centre Intelligence tool purely for auditing purposes. But Cirba believed its tool could also do capacity planning and analysis. "They took it away and came back with a quick tool that could do just that," says Bell senior systems analyst Lou Fachin.

Cirba claimed that its new tools could provide data centre intelligence: detailed reporting of asset utilisation in the data centre combined with cross-referenced information about what systems could be consolidated based on factors such as a server's operating system version, its utilisation percentage, its available memory, or seemingly trivial but often critical details such as the time zone setting of the system clock.

The tool generates reports that help users identify consolidation opportunities without resorting to extended whiteboard sessions or trial and error. "What I really like is the way they can set you up for consolidation," says Andi Mann, senior analyst at IT consultancy Enterprise Management Associates. "The Cirba stuff gives you some easy-to-use graphics and metrics on utilisation and compatibility. You could probably achieve some of this with some of the high-level management-type tools from IBM and BMC, but Cirba provides sort of a one-stop shop for this specific functionality."

Consolidation-specific tools aren't a necessity, however, particularly for smaller companies, says Michael Minichino, director of infrastructure at marketing services provider Parago. Heading into 2006, Minichino had several new IT initiatives slated that would require either more space at his co-location facility or better use of the existing racks.

Minichino was intrigued by the power-per-rack-space claims of new Sun hardware -- the T2000 series of servers -- and decided to try the latter route.

But he didn't bother looking for a tool to help him figure his savings; he went straight to spreadsheets. "We've purchased an asset management tool mainly for tracking workstations," Minichino says. "[But] I haven't really seen anything that would give me more of a return than spreadsheets." His calculations led him to cut 10 servers from his co-location facility.

Keep up online

For anyone looking to jump into the consolidation toolset on the cheap, Sun offers a free, downloadable Sim Data centre Java application that can calculate the power, heat and space requirements of your current data centre versus one with different hardware. (Sun supplies templates for its own and some competing systems, but you can also configure your own.)

Larger companies, however, may well see consolidation tools as a means of saving time and effort. As part of an application consolidation and tracking effort, David O'Neill, executive director of IT at Boise State University, started using a service discovery tool from startup software vendor nLayers.

Previously, identifying assets and connections between systems for audit or troubleshooting purposes meant making heavy demands on your systems engineers' energy. "You put your engineering staff at the whiteboard, give them a couple cans of pop, and they spend the rest of the day drawing pictures," O'Neill says.

With the nLayers tool, O'Neill is able to "let the machine do the inventory," and he can dedicate his engineers to more important tasks. NLayers also claims that its products can map the connections between systems and identify under-utilised servers.

If all these tools sound to you like features that should be part of larger-scale asset management, configuration management or business service management tools, BMC Software, IBM and other large vendors want to meet you.

BMC says it already has a suite of tools capable of initial device discovery, performance monitoring and analysis, configuration management and ongoing optimisation. What BMC's suite lacks, according to Dave Wagner, solutions management director for capacity management and provisioning at BMC, is an easy interface to tie all those operations together. But, he says, customers can expect to see the bigger vendors expand and improve their product lines, while the smaller vendors will consolidate or cooperate in order to provide the more wide-ranging management solution large corporations will need.

Who's being served?

No matter how insightful they may become about the technical configuration of your infrastructure, these tools will never be able to map your consolidation efforts to the political and contractual landscape of your corporation.

A word to the wise: Get in touch with the server owners well before you intend to absorb their beloved boxes and applications into your data centre. This will help smooth your path as well as help you identify relatively early on in the process if there are good reasons (compliance, security or otherwise) for keeping some seemingly underutilised hardware right where it is.

It's also worth noting that internal politics might be the least of your hurdles. "Quite often, [the difficulties lie in] vendor relations," says Bell Mobility's Tremblay. He notes that in one case Bell Mobility discovered a system consisting of 19 servers installed for one application by a certain vendor that Tremblay's team found could be consolidated to nine. Such inefficient installations are going to be a thing of the past, Tremblay says.

"We now say ‘no more,'" he insists, noting that Bell has created a policy of examining vendor architecture plans for efficiency before it agrees to an implementation. "[Vendors] have to start thinking of redesigning what they are selling" to make systems more efficient, Tremblay says.

If every IT organisation does the same, your next consolidation effort could be your last.