Predicting performance of a network order fulfillment system

Patent metadata

Field	Value
Publication number	US-11502894-B2
Application number	US-202017094491-A
Country	US
Kind code	B2
Filing date	Nov 10, 2020
Priority date	Nov 10, 2020
Publication date	Nov 15, 2022
Grant date	Nov 15, 2022

Abstract

A system may monitor transaction data pertaining to a plurality of transaction types received by a network order fulfillment system. The system may classify the transaction data into a plurality of alarm types based on pre-defined impact of an alarm type to a given transaction type. The system may analyze a plurality of performance parameters influencing a performance of the network order fulfillment system, and identify a performance parameter exhibiting an anomaly based on historical data, a current status of the plurality of the performance parameters and a predefined prediction model. The system may ascertain whether the identified performance parameter negatively impacts the performance of the network order fulfillment system, based on evaluation rules. The system may proactively implement a remediation action to remediate a potential fault caused by the identified performance parameter when the identified performance parameter negatively impacts the performance of the network order fulfillment system.

First claim

We claim:

1. A system comprising:
a processor;
an alarm analyzer coupled to the processor, the alarm analyzer to:
monitor transaction data received by a network order fulfillment system having a multi-platform configuration for fulfilling network tasks, the transaction data including transactions pertaining to a plurality of transaction types; and
based on monitoring, determine a real-time status of the transaction data to classify the transaction data into a plurality of alarm types, the classification being based on a pre-defined impact of an alarm type to a given transaction type;

a performance predictor coupled to the processor, the predictor to:
analyze a plurality of performance parameters corresponding to the transaction data influencing a performance of the network order fulfillment system;
determine a performance score for service applications of the network order fulfillment system, based on a weighted majority voting of the plurality of performance parameters, wherein the weighted majority voting is based on a decision tree for each of the plurality of the performance parameters;
identify a performance parameter exhibiting an anomaly, based on historical data, a current status of the plurality of the performance parameters, and a predefined prediction model when the performance score is less than a threshold score;
determine that the performance is being negatively impacted when the performance score is less than the threshold score; and
identify the performance parameter from the plurality of performance parameters exhibiting an anomaly, based on a corresponding alarm value; and

an auto remediator coupled to the processor, the auto remediator to:
ascertain whether the identified performance parameter negatively impacts the performance of the network order fulfillment system, based on evaluation rules;
proactively implement a remediation action to remediate a potential fault caused by the identified performance parameter when the identified performance parameter negatively impacts the performance of the network order fulfillment system.

2. The system as claimed in claim 1 , wherein the alarm analyzer is to:
analyze the transaction data to determine a volume of incoming transactions and a real-time status of each of the plurality of the performance parameters associated with the transaction data that is being monitored; and
provide the analyzed data to the performance predictor for predicting the performance of the network order fulfillment system.

3. The system as claimed in claim 1 , wherein the performance predictor is to implement at least one of a Random Forest technique, and a C4.5 technique combined with a regression based prediction approach.

4. The system as claimed in claim 3 , wherein the performance predictor is to:
implement the Random Forest technique to construct the predefined prediction model;
implement a linear regression technique to predict an alarm value for each of the plurality of the performance parameters; and
implement the C4.5 technique to generate a decision tree for each of the plurality of the performance parameters.

5. The system as claimed in claim 1 , wherein the auto remediator is to:
ascertain whether there are multiple performance parameters exhibiting anomalies;
correlate the multiple performance parameters to identify one or more performance parameters to take action on, based on the evaluation rules, when the multiple performance parameters exhibit the anomalies.

6. The system as claimed in claim 1 , wherein the plurality of performance parameters include an integration middleware node status, an integration middleware service status, a connection status of integration processing server with messaging queues, a software platform virtual machine status, a database status, a database query performance status, a database connectivity status, a performance of messaging threads between various technology platform components, a CPU utilization status, a memory utilization status, a response time for integrations between various platforms or applications within the platform, a connection status between an application and a database, a connection status between a middleware and applications, connection faults between network and a database, a mediation flow status and performance, data related issues stemming from the transaction data, a quality of data in the database, a performance of the middleware queues, and a middleware message transmission status.

7. The system as claimed in claim 1 , wherein the auto remediator is to implement a machine learning based evaluation model to evaluate a plurality of remediation action options available to remedy the potential fault.

8. A method comprising:
monitoring, by a processor, transaction data received by a network order fulfillment system having a multi-platform configuration for fulfilling network tasks, the transaction data including transactions pertaining to a plurality of transaction types;
based on monitoring, determining a real-time status of the transaction data to classify the transaction data into a plurality of alarm types, the classification being based on a pre-defined impact of an alarm type to a given transaction type;
analyzing, by the processor, a plurality of performance parameters corresponding to the transaction data influencing a performance of the network order fulfillment system;
determining, by the processor, a performance score for service applications of the network order fulfillment system, based on a weighted majority voting of the plurality of performance parameters, wherein the weighted majority voting is based on a decision tree for each of the plurality of the performance parameters;
identifying, by the processor, a performance parameter exhibiting an anomaly, based on historical data, a current status of the plurality of the performance parameters, and a predefined prediction model when performance score is less than the threshold score;
determining, by the processor, that the performance is being negatively impacted when the performance score is less than the threshold score;
identifying, by the processor, the performance parameter from the plurality of performance parameters exhibiting the anomaly, based on a corresponding alarm value;
ascertaining, by the processor, whether the identified performance parameter negatively impacts the performance of the network order fulfillment system, based on evaluation rules; and
implementing proactively, by the processor, a remediation action to remediate a potential error caused by the identified performance parameter when the identified performance parameter negatively impacts the performance of the network order fulfillment system.

9. The method as claimed in claim 8 , further comprising:
analyzing, by the processor, the transaction data to determine a volume of incoming transactions and a real-time status of each of the plurality of the performance parameters associated with the transaction data that is being monitored; and
providing, by the processor, the analyzed data for predicting the performance of the network order fulfillment system.

10. The method as claimed in claim 8 , wherein identifying the performance parameter exhibiting the anomaly includes implementing, by the processor, at least one of a Random Forest technique, and a C4.5 technique combined with a regression based prediction approach.

11. The method as claimed in claim 10 , further comprising:
implementing, by the processor, the Random Forest technique to construct the predefined prediction model;
implementing, by the processor, a linear regression technique to predict an alarm value for each of the plurality of the performance p

Assignees

• Accenture Global Solutions Ltd

Inventors

• Mohan Mahesh Kumar
• Kumar Madan
• Kumar Harshit

Classifications

• H04L41/5025

  by proactively reacting to service quality change, e.g. by reconfiguration after service quality degradation or upgrade · CPC title

• H04L43/08

  Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters · CPC title

• H04L41/0886

  Fully automatic configuration · CPC title

• H04L41/0816

  the condition being an adaptation, e.g. in response to network events · CPC title

• H04L41/5054

  Automatic deployment of services triggered by the service manager, e.g. service implementation by automatic configuration of network components · CPC title