Browse Prior Art Database

Technique for Managing Network Counters

IP.com Disclosure Number: IPCOM000102392D
Original Publication Date: 1990-Nov-01
Included in the Prior Art Database: 2005-Mar-17
Document File: 4 page(s) / 129K

Publishing Venue

IBM

Related People

Irvin, DR: AUTHOR [+2]

Abstract

Keeping various sets of counters has become a key aspect of managing communication networks. Such counters monitor the operation of the communications function, and detect degradation, impending failure, and failure of various network components and links.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Technique for Managing Network Counters

       Keeping various sets of counters has become a key aspect
of managing communication networks.  Such counters monitor the
operation of the communications function, and detect degradation,
impending failure, and failure of various network components and
links.

      Having knowledge of the network's statistics is only the first
step in ensuring that the network meets established goals for
performance and availability.  A second aspect of the overall problem
involves examining the data gathered by the counters and deciding if
the network is operating properly.  When a network element fails, or
gives signs of an impending failure, an operator should be notified
so that corrective action may begin.  In the data processing world
such notifications are known as "alerts"; in the world of telephony
they are called "alarms"  This invention addresses the problem of
monitoring the set of network counters and deciding when an alert or
alarm should be issued.

      A simple example of a set of counters for a hypothetical data
link control (DLC) would be the set consisting of the numbers of
frames received with error, frames sent, and frames sent with error.
When some predetermined criterion is met, the physical unit (PU)
maintaining the counter set sends the accumulated data to the control
point that owns the link in question.  From this data, the ratio of
errors-to- traffic (E/T) is computed.  Comparing the computed value
of E/T to predetermined thresholds provides a means of categorizing
the operation of the link as normal, degraded or failed.  In a
non-hierarchical system, the responsible PU and the owning control
point may be resident in the same box, or the PU may itself perform
the computation of E/T.

      Two methods are presently used to manage the counters. In the
first method, E/T is computed when the error-count exceeds a
pre-established threshold.  In the second method, E/T is computed
when a predetermined measurement period times out.

      A new method of managing counter thresholds, as shown in the
flowchart in Fig. 1, will now be described.  In Fig. 1, parameter "T"
is the duration of the measurement period; a given value of T applies
to a set of management counters. Parameter "N" is a threshold for a
particular counter within the set.  Thus, the management of a set of
counters involves a common measurement period of duration T, and a
set of thresholds, one for each counter, where the particular value
of the threshold for the counter shown in the flowchart is N.
Parameter "t" measures the flow of time, and is kept with modulo T;
thus, t advances from 0 to T, and is reset to 0 on each occurrence
of T.  Parameter "n" indicates...