May 11, 2001, 11:57 AM — At the foundation of today's networks lies a fragile, interdependent and sophisticated ecosystem of hardware and software configurations. Unfortunately this complexity has a downside. Problems arising from system misconfigurations are the largest source of network failures, system downtime, help desk calls and security breaches.
A new management discipline called configuration management is emerging to proactively manage the millions of configuration settings that exist in a typical enterprise network.
In an effort to ensure the highest levels of reliability for their networks, most companies establish standard configuration builds for their servers and workstations. But once these machines are deployed in production environments, their configurations inevitably begin to change.
Some modifications are planned, such as updates to drivers and changes in security policy, while others are unplanned, such as quick adjustments made by administrators to fix problems. Configuration management involves managing and monitoring these configuration standards. With centralized access to this critical information, companies can gain control over their networks.
The only way to track the enormous amount of configuration data is to deploy an intelligent agent on each device in the network. In large environments, a (one machine collecting from multiple devices) won't scale. Furthermore, the pain associated with managing the proxy relationships overrides the benefits.
Using the agent model, detailed configuration data such as device drivers, disk configuration and application data can be collected and transmitted to a central management repository.
The configuration management architecture should provide the flexibility to schedule data collections at frequent intervals as well as the ability to perform on-request collections. The agent should only transmit data that has changed since the previous collection.
The management repository, usually a database, provides access to all of the configuration data in one centralized location. This database maintains detailed information on configuration changes made to each machine. With all of this data in a central database, including current machine configurations and a log of changes, you can get immediate answers to complex questions about a network.
The central repository provides instant visibility into how machines are configured, and makes it possible to efficiently manage configuration changes across thousands of machines.
Due to the volume of configuration data that exists in a typical company, analytical tools are required to extract usable information from the database and let administrators: