# US 2004/0204906 A1 · Method and apparatus for event monitoring in an information processing system

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2004/0204906 A1
- Document ID: US20040204906A1
- Status: Published application
- Kind: A1
- Application number: 10/811,291
- Filed: 2004-03-26
- Published: 2004-10-14
- Earliest claimed: 2003-03-31
- Assignees: Lucent Technologies Inc.
- Inventors: Diane Lambert; Chuanhai Liu
- Web page: https://patentgrep.com/patent/US20040204906A1

## CPC

- H04L 43/00: Arrangements for monitoring or testing data switching networks
- H04M 3/36: Statistical metering, e.g. recording occasions when traffic exceeds capacity of trunks
- H04L 43/106: using time related information in packets, e.g. by adding timestamps
- H04L 43/16: Threshold monitoring
- H04M 2201/12: Counting circuits
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- H04M: TELEPHONIC COMMUNICATION (circuits for controlling other apparatus via a telephone cable and not involving telephone switching apparatus G08)

## Abstract

In an event monitor associated with an information processing system, data values corresponding to monitored events are processed using a set of parameters characterizing a reference distribution of the data values. From a given one of the data values a corresponding probability value is computed utilizing the set of parameters. A thresholding operation is then performed on the probability value, and alarm information is generated based on the result of the thresholding operation. The set of parameters are periodically updated to take into account one or more of the data values.

## Claims

### Claim 1

A method for monitoring data values in an event monitoring system, using a set of parameters characterizing a reference distribution of the data values, the method comprising the steps of:
computing from a given one of the data values a corresponding probability value, utilizing the set of parameters characterizing the reference distribution;
performing a thresholding operation on the probability value;
generating alarm information based on the result of the thresholding operation; and
periodically updating the set of parameters to take into account one or more of the data values.

### Claim 2 (depends on 1)

The method of claim 1 wherein the event monitoring system comprises an event counter which generates the data values and an event monitor which processes the data values to generate the alarm information.

### Claim 3 (depends on 1)

The method of claim 1 wherein the set of parameters comprises at least a mean and a variance for the reference distribution of the data values.

### Claim 4 (depends on 3)

The method of claim 3 wherein the reference distribution comprises an asymmetric distribution.

### Claim 5 (depends on 4)

The method of claim 4 wherein the asymmetric distribution comprises one of a negative binomial distribution, a lognormal distribution and a Poisson distribution.

### Claim 6 (depends on 1)

The method of claim 1 wherein the data values are collected and thresholded in accordance with a measurement timescale, the set of parameters are updated in accordance with a parameter timescale longer in duration than the measurement timescale, and a complete set of the parameters is determinable in accordance with a cycle timescale longer in duration than the parameter timescale.

### Claim 7 (depends on 6)

The method of claim 6 wherein the measurement timescale comprises approximately one minute, the parameter timescale comprises approximately one hour and the cycle timescale comprises approximately one day.

### Claim 8 (depends on 1)

The method of claim 1 wherein the probability value computed from a given data value x comprises at least one of:
(i) an upper tail probability P[data value≧x];
(ii) a lower tail probability P[data value≦x]; and
(iii) a minimum of {P[data value≧x], P[data value≦x]};
wherein the probability value is computed using the set of parameters characterizing the reference distribution that applies at the time the given data value x is collected.

### Claim 9 (depends on 1)

The method of claim 1 wherein the set of parameters characterizing the reference distribution is determined at least in part by interpolating coefficients over multiple parameter timescales.

### Claim 10 (depends on 1)

The method of claim 1 further comprising the step of periodically monitoring validity of the reference distribution.

### Claim 11 (depends on 10)

The method of claim 10 wherein the validity of the reference distribution is monitored by generating a histogram of a plurality of probability values and determining if counts associated with a plurality of intervals of the histogram are each approximately the same.

### Claim 12 (depends on 11)

The method of claim 11 wherein the probability values utilized in generating the histogram are corrected for continuity by:
computing the probability of observing data value x under the reference distribution as px=pL+pU−1,where pL and pU denote respective lower and upper tail probabilities associated with data value x under the reference distribution;
taking a random draw Z from a uniform distribution on [0, 1]; and
computing a continuity corrected probability value as pcont=pL−Z*px.

### Claim 13 (depends on 1)

The method of claim 1 further comprising the step of determining if the given data value is an outlier, and if so modifying the given data value as follows:
replacing an upper tail outlier with a random draw from the reference distribution conditioned to be within a specified upper portion of the distribution, such that the upper tail outlier is replaced with another upper tail data value; and
replacing a lower tail outlier with a random draw from the reference distribution conditioned to be within a specified lower portion of the distribution, such that the lower tail outlier is replaced with another lower tail data value.

### Claim 14 (depends on 1)

The method of claim 1 wherein the step of updating the set of parameters characterizing the reference distribution further comprises updating a given one of the parameters as a weighted average of corresponding parameter values determined over a designated timescale.

### Claim 15 (depends on 1)

The method of claim 1 further including the step of initializing the set of parameters characterizing the reference distribution.

### Claim 16 (depends on 1)

The method of claim 1 wherein the alarm information further comprises an alarm severity measure that indicates the severity of alarm conditions based on multiple-mode error comprising at least error duration and error spread.

### Claim 17 (depends on 1)

The method of claim 1 wherein the alarm information further comprises an alarm severity measure generated at least in part utilizing at least one of a ratio of multiple probability values and a difference of logarithms of probability values.

### Claim 18 (depends on 1)

The method of claim 1 wherein the alarm information further comprises an alarm severity measure S, and the given data value has a logit transformation L, for a corresponding probability value p, given by:
3 L ⁡ ( p ) = log ⁡ ( 1 - p p ) ,
and further wherein the alarm severity measure is updated as follows:
Snew=(1−w)*S+w*L,
where w denotes a severity weight.

### Claim 19

An apparatus for monitoring data values in an event monitoring system, using a set of parameters characterizing a reference distribution of the data values, the apparatus comprising:
a memory for storing the set of parameters; and
a processor coupled to the memory and operative to control operations associated with the monitoring of the data values, the operations including:
computing from a given one of the data values a corresponding probability value, utilizing the set of parameters characterizing the reference distribution;
performing a thresholding operation on the probability value;
generating alarm information based on the result of the thresholding operation; and
periodically updating the set of parameters to take into account one or more of the data values.

### Claim 20

An article of manufacture comprising a machine-readable storage medium for storing one or more programs for use in monitoring data values in an event monitoring system, using a set of parameters characterizing a reference distribution of the data values, wherein the one or more programs when executed implement the steps of:
computing from a given one of the data values a corresponding probability value, utilizing the set of parameters characterizing the reference distribution;
performing a thresholding operation on the probability value;
generating alarm information based on the result of the thresholding operation; and
periodically updating the set of parameters to take into account one or more of the data values.
