Rational Strategies for Multiple Tool Monitoring

As the years go by, IT services monitoring and management tools tend to multiply on an ad-hoc basis. Many enterprises find themselves burdened with these systems, rather than supported by them. The original purpose of the software -- infrastructure availability, performance, incident response and consistent management information -- can sometimes become lost in conflicting and non-interoperable systems or multiple information sources.

"This proliferation of IT management tools results from many factors including turnover of management and key technical personnel, persuasive software vendors, and the cumulative effect of mergers, acquisitions, and divestments," says Dave Lilly, General Partner at CapTech, providers of pre-packaged open source IT monitoring solutions and consulting.

IT managers of one large U.S. bank, for example, identified 180 separate purchases of monitoring tools over the last several years including one product that had been purchased from four different vendors. While perhaps an extreme example, this scenario is far from uncommon.

According to U.K.-based IT analysts Conspectus, "Overall, the picture is one where companies understand the difficulties they face in managing their IT infrastructure successfully, but are still at an early stage in finding the best ways to achieve this."

In reality, there are two potential ways to resolve this dilemma -- replacing the accumulated monitoring and management tools wholesale with a single enterprise class system such as OpenView, Patrol, Tivoli, Net Cool, etc, known as framework solutions; or alternatively, developing systems architecture for the integration of individual tools to create your own solution.

Enterprise Class Tools: Typically, these offer broad system monitoring and management functionality, including software distribution, asset management and configuration as well as application management; across the spectrum of hardware and software platforms. They tend to be a collection of individual solutions that have been acquired to create a single offering, with an over-arching integration layer. A good example is IBM's purchase of Think Dynamics to bolster Tivoli in the area of dynamically reallocating systems resources.

Frameworks are particularly suitable for large enterprises where the cost of software, implementation and maintenance is diluted within the sheer extent of the network. However, they may not provide the best capability for each of the various functionalities required. The most suitable management tools for a particular platform or technology tend to be provided by the vendor -- CiscoWorks for Cisco hardware, Enterprise Manager for Oracle, etc.

The Multi Tool Environment: A multi-tool approach potentially offers a "best-of-both-worlds" scenario.

"An alternative approach to the across-the-board deployment of a single framework solution is to develop an architecture and integration strategy for the simultaneous use of multiple tools," said Lilly. "To accomplish this, monitoring tools are assigned different functional responsibilities that are referred to as 'Levels.'"

Level 1 tools monitor availability of applications and infrastructure components and respond to failures by notification and escalation. Level 1 monitoring can be accomplished by open source tools such as Nagios or Mon or with inexpensive software like Big Brother.

Level 2 tools are those that provide warnings of impending failures and notification of the need to perform situational maintenance, or which collect data for use in capacity management. Usually Level 2 tools also perform configuration/asset management and software distribution functions for the components that they manage. These tools are typically those provided by hardware and software vendors to monitor and manage their own system.

To implement the rationalized multi-tool approach, a single tool is chosen as the Level 1 system across the entire infrastructure. Thus, each component of the infrastructure, whether hardware or software, is monitored both by the relatively simple Level One monitoring tool and a specific Level Two monitoring and system management tool.

An important factor to take into account, though, is the degree of existing standardization. If all network hardware is sourced from a single vendor, all DBMS from a single source, and so on, those vendors' management products can be utilized at Level 2 to create a unified system.

If multiple vendors systems have been integrated over time, then greater the complexity will be present. The key here is to take whatever customized solutions have been built and wrap these around a Level 1 tool to provide a single management console. Large organizations will generally already own a license to a simple Level 1 monitoring tool that is not being fully utilized.

Alternatively, low-cost open source tools exist. Lilly estimates that the savings for a rationalized multi-tool approach is typically 50% to 75% over a framework implementation (For more information, see White Paper -- Rationalizing The Multi-Tool Monitoring Environment.

Implementation

In broad terms, a multi-tool implementation is a four-step process. As with all enterprise deployments, a full understanding of the existing IT service management business processes must be in place. With that, you can overlay a map of Level 1 and Level 2 systems and define in detail what each of those is going to achieve, how they will integrate with each other, what solutions are best-fit, what gaps need filling, and so on.

The first stage of deployment is the Level 1 tool, incorporating infrastructure failure alarms, escalation and notification processes, availability and responsiveness reports, and IT dashboards. If an on-line ticketing system is in place, the Level 1 tool can be integrated into this. Level 2 tools are implemented one at a time, integrating their alarm messaging into the Level 1 displays, dashboards and reports.

The final step is integration with the Help Desk system and deployment of the complete IT Service Management dashboard and reporting.

End of the Road

Within a few years, the choice between these various methods of IT service management may disappear. A new breed of tools are on the immediate horizon which threaten to change the game entirely.

"Today's systems management tools are at the end of their road," said Laura Koetzle at Forrester Research. "New organic IT management software will replace them, eventually culminating in robust, self-healing infrastructure by 2008."

Until then, organizations have the choice of framework versus a multi-tool approach.