A `Control-System` is the design of closed-loop or feedback control systems. Control Systems are in fact studied extensively in other engineering disciplines such as mechanical and aeronautical engineering; In these engineering disciplines, designers most often employ linear control theory, which uses input-output relationships of linear systems to study controller properties such as: stability (finite inputs produces finite outputs), bias (how well objectives are achieved), rise time (how quickly the system responds to a change in objective), and settling time (how long until steady state is reached). This theory provides sound and rigorous mathematical principles for the design and analysis of closed-loop systems.
A `Network Management Control-System` is a Closed-loop or feedback management control system applied to NMS systems through the management processes to generate close-loop management solutions that reports on stability, bias, rise time, settling time of the overall or parts of the network or a zone within the network being managed.
Now; to effectively build on this concept lets classify the network management activity levels and from within those levels we can try to understand how `Network Management Control-Systems` could be structured.
There are four levels of activity (Feed-back activities) that one must understand before applying management to a specific service or device. These four levels of activity are as follows:
This is the case when no monitoring is being done and if you did receive an alarm in this area, you would ignore it.
This is where you react to a problem after it has occurred yet no monitoring has been applied.
This is where you are monitoring components but must interactively troubleshoot to eliminate the side effect alarms and isolate to a root cause. Later on you could test for the symptoms early on and practively..
This is where you are monitoring components and the system provides a root cause alarm for the problem at hand and automatic restoral processes are in place where possible to minimize downtime.
This new concept that we are referring to is what we will call `Control Management`. The goals of Control management as we see it at this point are:
To interactively troubleshoot and eliminate side effect alarms.
To isolate alarms to a root cause.
To trigger automatic restoral processes whenever a root cause alarm is detected.
To measure input-output relationship performance metrics like:
- Stability (finite inputs produces finite outputs),
- Bias (how well objectives are achieved),
- Rise time (how quickly the system responds to a change in objective), and
- Settling time (how long until steady state is reached).
These four levels of activities outline exactly how a typical support organization is dealing with problems today and where any network administrator want them to be in terms of goals. Within the support organization are teams with different goals and focuses (i.e. Unix support, desktop support, network support, etc.). its essential to remember that while a specific alarm may warrant an inactive approach by one team, to another team it may demand a proactive approach. These goals should be kept in mind when gathering requirements for network management.