Communications: electrical – Continuously variable indicating – With meter reading
Patent
1995-08-29
1998-05-05
Horabik, Michael
Communications: electrical
Continuously variable indicating
With meter reading
34082507, 34082536, 34082506, 370242, 370245, 39520011, 395180, G06F 15177, H04Q 100
Patent
active
057480982
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to the operation of telecommunications network management systems.
2. Related Art
These systems are used for monitoring, and in some cases also controlling, the operating states of interacting network elements found in the communications networks. In a network management system, monitoring devices sense the operating states of the various network elements and send signals to a network manager in response to significant events in the operation of their respective equipments. Signals may also be sent by human agency to report conditions in the equipment. The term "event" will be used in this specification to signify any state, or change in state, causing such a signal to be sent. In general these events will be faults and the monitoring devices will be fault detectors for sending alarm signals to the network management system in the event of a fault being detected.
For example, in a telecommunications network, an alarm signal caused by the failure of a switching centre would alert the system manager who would arrange for alternative routings to be made and who would also arrange for any necessary repair work to be done on the faulty switching centre. In some systems these responses may be automated, but more usually the faults will require human intervention, the system merely providing to the system manager details of any faults requiring attention. This allows the manager to organise the available resources efficiently, taking into account factors such as safety-criticality, priority, and the whereabouts of field staff.
This also allows non-significant alarms with known causes, such as those caused by equipment having been disconnected for routine maintenance, to be disregarded by the system manager.
It is well established that effects of a network failure, such as the failure of a high bit rate line system, will propagate down through a hierarchy of dependent resources and initiate many nearly simultaneous alarm messages. Time and other resources can be wasted by investigating the sources of all the alarms if the underlying cause has not been identified.
A fault may affect all equipment directly connected to the source of the fault, or all equipment at one geographic location, (although they may be topologically remote from each other) or all equipment of a specific type. For example, external radio interference could affect all radio links in the network operating on a specific frequency within radio range of the source of interference although, from the point of view of connectivity of the network, they may appear remote from each other. Based upon this understanding it is known that alarm messages that occur with close temporal proximity will tend to be associated or correlated.
The size of any temporal window within which alarm messages are considered as correlated has to be optimised--if it is set too large, the chance of alarm messages from unconnected resources arriving within the temporal window increases; if it is set too small, only a subset of the dependent alarm messages might arrive within the temporal window.
In certain circumstances, related faults may only be detected by the system at a later time. For example, this will be the case if an equipment not in operation at the time of the initial fault later attempts to establish contact with the failed equipment.
The operator of the network will be able to determine an appropriate window size according to the characteristics of the network. This will depend on the nature of the network and the faults being monitored within it. In circumstances where faults are reported to the system by human agency rather than automatically, or in which their exact timing is difficult to measure, the window may be of the order of hours, or even days, whereas for systems continuously and automatically monitored a suitable time window may be measured in seconds.
Of course, the mere fact that two events have occurred together on one occasion is not necessarily indicative that there is a
REFERENCES:
patent: 4550278 (1985-10-01), Yamanaka
patent: 5155480 (1992-10-01), Pfeiffer
patent: 5265103 (1993-11-01), Brightwell
patent: 5309448 (1994-05-01), Bouloutas et al.
patent: 5334970 (1994-08-01), Bailey
patent: 5528759 (1996-06-01), Moore
Beaulieu Yonel
British Telecommunications public limited company
Horabik Michael
LandOfFree
Event correlation does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Event correlation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Event correlation will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-58707