Data processing system, data processing method, and program

Patent metadata

Field	Value
Publication number	US-9417981-B2
Application number	US-201114110666-A
Country	US
Kind code	B2
Filing date	May 11, 2011
Priority date	May 11, 2011
Publication date	Aug 16, 2016
Grant date	Aug 16, 2016

Abstract

A data processing system for storing observation data including an observation value corresponding to a point in a storage area and calculating a comparison value to be compared with a future observation value, the data processing system comprising: an observation data collection unit for specifying first observation data and second observation data which are observation data before the first point; a past similar data selection unit for selecting parts for the same duration as the first predetermined time, the observation values of which are identical to the observation values of the first observation data, from the second observation data as past similar data; a parameter data selection unit for selecting the observation values for a predetermined time from the latest time of the parts selected as the past similar data as parameter data; and a comparison value calculation unit for calculating the comparison value based on the parameter data.

First claim

What is claimed is:

1. A data processing system for storing observation data including an observation value corresponding to a point of an observation target in a storage area and calculating a comparison value to be compared with a future observation value based on observation data stored in the storage area, the statistical value calculation system comprising:
a hardware observation data collection unit for periodically collecting the observation data from the observation target, storing the collected observation data in the storage area and specifying first observation data from a first point earlier by a first predetermined time than a current point to the current point and second observation data which are observation data before the first point;
a hardware past data selection unit for selecting at least one of parts for the same duration as the first predetermined time, the observation values of which are identical or within a predetermined range to the observation values of the first observation data for the first predetermined time, from the second observation data as past data within a predetermined range;
a hardware parameter data selection unit for selecting the observation values for a second predetermined time from the latest time of at least one of the parts selected as the past data within a predetermined range as parameter data;
a hardware comparison value calculation unit for calculating the comparison value based on the parameter data; and
a hardware comparison value determination unit for determining whether or not an observation value of observation data newly collected by the hardware observation data collection unit is in excess of the comparison value, wherein the hardware comparison value determination unit:
displays the effect that the observation value is in excess of the comparison value, in a case where the observation value of the newly collected observation data is determined to be in excess of the comparison value; and
displays a part, in which the observation value is in excess of the comparison value, in a color different from the color of a part in which the observation value is not in excess of the comparison value;

wherein the hardware past similar data selection unit:
calculates a characteristic amount indicating a characteristic of the observation values of a part for the same duration as the first predetermined time of the observation data;
selects at least one of parts of the second observation data, which at least one of parts is a characteristic amount in a predetermined range from a characteristic amount of the first observation data, as a past data candidate;
calculates a value of all the observation values of each selected past data candidate with respect to the observation values included in the first observation data; and
selects the past data candidate identical or similar to the first observation data as the past data based on the calculated value;

wherein the observation target is data on the operation of a computer system;
the observation data includes process information capable of specifying a process performed by the computer system at a point corresponding to the observation value; and
the hardware past data selection unit manages day-of-week information capable of specifying the day of week of observation of the observation data and the process information together with the characteristic amount.

2. The data processing system according to claim 1 , wherein:
the hardware parameter data selection unit selects a plurality of pieces of the parameter data;
the hardware comparison value calculation unit calculates the comparison value based on the observation values of a utilization part which is a part of each pieces of the parameter data and includes a hardware utilization part specification unit for specifying the utilization part from each pieces of the parameter data;
the hardware utilization part specification unit:
calculates an average value of the observation values of all the pieces of selected parameter data at each point of time and calculates a variance of the observation values at each point from the average value;
specifies a part of each pieces of the parameter data from an earliest point to a point, at which the variance becomes equal to or higher than a predetermined value, as the utilization part;

the hardware comparison value calculation unit relates an earliest point of the utilization part specified by the utilization part specification unit to the current point and calculates the comparison value for time of the utilization part after the current point; and
the hardware past similar data selection unit selects new past data, in a case where the time of the utilization part elapses.

3. The data processing system according to claim 1 , wherein the hardware comparison value calculation unit:
calculates an average value and a standard deviation of the observation values of all the pieces of parameter data selected by the parameter data selection unit at each point;
calculates a multiplication value obtained by multiplying the standard deviation by a given parameter α; and
sets a value obtained by adding the multiplication value to the calculated average value as an upper limit value of the comparison value and sets a value obtained by subtracting the multiplication value from the calculated average value as a lower limit value of the comparison value.

4. The data processing system according to claim 1 , wherein the hardware past similar data selection unit:
calculates an average value of the observation values of the part of the observation data for the same duration as the first predetermined time or an average value of gradients of the observation values of the part as the characteristic amount;
calculates a coefficient of correlation between the observation values included in the first observation data and the observation values included in each selected past similar data candidate as the calculated value; and
selects the past similar data from the past similar data candidates based on values of the calculated coefficients of correlation.

5. A data processing method for calculating a comparison value to be compared with a future observation value based on observation data stored in a storage area in a computer system for storing observation data including an observation value corresponding to a point of an observation target in the storage area, the method including:
an observation data collection step including a step of periodically collecting the observation data from the observation target, a step of storing the collected observation data in the storage area and a step of specifying first observation data from a first point earlier by a first predetermined time than a current point to the current point and second observation data which are observation data before the first point;
a past data selection step of selecting at least one of parts for the same duration as the first predetermined time, the observation values of which are identical or within a predetermined range to the observation values of the first observation data for the first predetermined time, from the second observation data as past data, wherein the past data selection step includes:
a step of calculating a characteristic amount indicating a characteristic of the observation values of a part for the same duration as the first predetermined time of the observation data;
a step of selecting at least one of parts of the second observation data, which at least one of parts is a characteristic amount lying in a predetermined range from a characteristic amount of the first observation data, as a past data candidate;
a step of calculating a value of all the observation values of each selected past data candidate with respect to the observation valu

Assignees

• Murakami Junichi
• Hitachi Ltd

Inventors

• Murakami Junichi

Classifications

• G06F11/3003 Primary

  Monitoring arrangements specially adapted to the computing system or computing system component being monitored · CPC title

• H04L41/142 Primary

  using statistical or mathematical methods · CPC title

• H04L41/147

  for predicting network behaviour · CPC title

• H04L43/067

  using time frame reporting · CPC title