The OSI has published an official definition of network management that includes five different components: configuration management, fault management, performance management, security management, and accounting management. I usually think of performance management as being composed of two separate subcomponents. The first is a tactical performance management, and the second is the more strategic long-term capacity planning component.
Configuration management actually includes two different but related activities. The first keeps track of physical hardware, serial numbers, locations, patching information, and so forth. The second part of configuration management is the process of modifying, backing up, and restoring the software configuration of network equipment. This aspect of configuration management often becomes the focus of the whole activity. Many hardware vendors for routers and switches have excellent software for building and modifying software configurations. This software usually includes the ability to do scheduled backups of running configurations. This ability is an extremely important feature. If you have a recent configuration backup, then replacing a failed router with a new one is a fast and easy operation. Without a backup, this replacement is time consuming and usually requires an experienced engineer to reconstruct the software configuration.
However, remember the physical tracking side of configuration management, especially if you deal with the configurations of Layer 2 devices such as hubs and switches. If network managers have accurate information about physical locations, MAC addresses, and cabling for end devices such as user workstations, then they can easily handle hardware moves, adds, and changes. In most organizations, business requirements force network administration to respond quickly and efficiently to requests for end-user moves and service changes. However, the cabling and hardware records are usually out-of-date, so every small move requires a technician to visit the site and carefully document the equipment and cabling. This process is expensive and slow.
Unfortunately, no software can solve this problem; it is primarily a procedural issue. Technicians making changes have to keep the records up-to-date, and the cabling and patch panels have to be periodically audited to ensure accuracy of the records. However, the network designer can do much to facilitate this process. If the patch panels are well designed and there is a clear correlation between physical floor location and cable numbers, then the technicians can at least get a running start at the job.
Fault management is what most people picture regarding network management. This management is the active monitoring of the various key network components to find problems and alert the appropriate people. But there is another side to fault management that is also important, particularly to the network designer—the troubleshooting process.
Troubleshooting occurs after the appropriate person knows of a problem. Usually, all that the fault-management software says is that a failure occurred somewhere near a particular device. It is usually not able to say what caused the problem, precisely which device needs attention, or even what the failure actually was. Upon receiving an alert, the network engineer must troubleshoot the problem, try to isolate the source, and look for a solution. For many problems there is a short-term solution to get the network back up immediately, as well as a long-term solution to make sure it doesn't happen again.
Performance management requires monitoring the network carefully and looking for bottlenecks and congestion issues. There is some overlap between performance management and fault management when performance problems become so severe that they interfere with the basic functioning of the network.
Capacity planning is the natural outcome of performance management. When network managers discover a systematic performance problem, such as a bandwidth shortage through performance management, they turn to capacity planning to resolve this problem. Capacity planning is fundamentally a network-design issue.
Security management is the set of activities that ensure that the network's security measures work properly. Every firewall must be carefully monitored to see if it is in danger of compromise or if it is being abused in some way. Similarly, security management includes the maintenance of any filtering or encryption options.
Security management leads directly into the concept of accounting management. Accounting partly deals with security. One of the main reasons for giving individual users different accounts is to ensure that they can only have access to the resources they require. This access is essentially a security issue. However, accounting management also includes the general problem of keeping track of who uses what on the network. In some cases, this information is used to bill for these services.
It should now be clear that all of the different activities of network management have network-design implications.