Predicting Network Performance

• US 20170034720A1
• Filed: 07/28/2015
• Published: 02/02/2017
• Est. Priority Date: 07/28/2015
• Status: Abandoned Application

• Current Assignee: Futurewei Technologies Incorporated (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Futurewei Technologies Incorporated (Huawei Investment & Holding Co., Ltd.)
• Inventors: Mathew, James, Ren, Yong, Gopalakrishnan, Nandu, Yang, Jin, Roa, Juan, Sheen, Baoling S.
• Application Number:
  US14/810,699
• Publication Number:
  US 20170034720A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: H04W 16/18 Network planning tools
• CPC Class Codes: H04W 16/22 Traffic simulation tools or models
• CPC Class Codes: H04W 24/04 Arrangements for maintaining operational condition
• CPC Class Codes: H04W 24/08 Testing, supervising or monitoring using real traffic
• CPC Class Codes: H04W 24/10 Scheduling measurement reports ; Arrangements for measurement reports

Abstract

Methods and systems for predicting network performance include receiving a number of sets of data points of a number of network elements. Each of the number of sets of data points includes performance counter values and a performance indicator of a respective network element of the number of network elements. A global model representing a global relationship pattern between the performance indicator and the performance counter values is determined based on the number of sets of data points of the number of network elements. For each network element, residual features are determined based on error measures between the global model and the set of data points including the performance indicator and the performance counter values of the network element. The number of network elements are clustered into a number of clusters based on the determined residual features of the number of network elements.

Claims

1. A method comprising:

receiving, by operation of a processing apparatus, a plurality of sets of data points of a plurality of network elements, each of the plurality of sets of data points corresponding to a respective network element of the plurality of network elements, the set of data points comprising performance counter values and a performance indicator of the respective network element;

determining, by operation of the processing apparatus, a global model representing a global relationship pattern between the performance indicator and the performance counter values based on the plurality of sets of data points of the plurality of network elements;

for each network element of the plurality of network elements, determining, by operation of the processing apparatus, one or more residual features, the one or more residual features based on error measures between the global model and the set of data points comprising the performance indicator and the performance counter values of the network element; and

clustering, by operation of the processing apparatus, the plurality of network elements into a plurality of clusters based on the determined one or more residual features of the plurality of network elements.

2. The method of claim 1, wherein the performance counter values comprise one or more of a number of active users in the network counter, a number of traffic bytes in the network, a throughput of the network, an interference level, or a downlink (DL) transmit power level.

3. The method of claim 1, wherein determining a global model representing a global relationship pattern between the performance indicator and the performance counter values based on the plurality of sets of data points of the plurality of network elements comprises performing a regression based on the plurality of sets of data points of the plurality of network elements.

4. The method of claim 1, wherein clustering the plurality of network elements into a plurality of clusters based on the determined one or more residual features of the plurality of network elements comprises clustering the plurality of network elements into a plurality of clusters without user equipment (UE) measurement reports (MRs), call history records (CHRs), configuration parameters, or engineering parameters of the plurality of network elements.

5. The method of claim 1, further comprising, prior to determining the global model, performing an additional layer of clustering by clustering the plurality of network elements into a plurality of super-clusters based on one or more features, and

wherein determining a global model comprises determining a global model for each of the plurality of super-clusters.

6. The method of claim 5, wherein the one or more features are determined based on the performance counter values and represent coverage, interference, or user equipment characteristics of the plurality of network elements.

7. The method of claim 5, wherein the one or more features comprises a correlation between an interference measurement and a traffic measurement of a network element or comprises a correlation between a call drop rate and an interference measurement of a network element.

8. The method of claim 1, further comprising, for each of the plurality of clusters, determining a respective regression model based on performance counter values and performance indicators of network elements within the cluster.

9. The method of claim 8, further comprising predicting performance of a network element according to the regression model.

10. The method of claim 8, further comprising selecting cell physics features that are indicative of network performance behavior based on the plurality of clusters and the respective regression models.

11. A computing system comprising:

a memory storing programming; and

a processor interoperably coupled with the memory and, when executing the programming, the computing system is configured to;

receive a plurality of sets of data points of a plurality of network elements, each of the plurality of sets of data points corresponding to a respective network element of the plurality of network elements, the set of data points comprising performance counter values and a performance indicator of the respective network element;

determine a global model representing a global relationship pattern between the performance indicator and the performance counter values based on the plurality of sets of data points of the plurality of network elements;

for each network element of the plurality of network elements, determine one or more residual features, the one or more residual features based on error measures between the global model and the set of data points comprising the performance indicator and the performance counter values of the network element; and

cluster the plurality of network elements into a plurality of clusters based on the determined one or more residual features of the plurality of network elements.

12. The computing system of claim 11, wherein the performance counter values comprise one or more of a number of active users in the network, a number of traffic bytes in the network, a throughput of the network, an interference level, or a downlink (DL) transmit power level.

13. The computing system of claim 11, wherein determining a global model representing a global relationship pattern between the performance indicator and the performance counter values based on the plurality of sets of data points of the plurality of network elements comprises performing a regression based on the plurality of sets of data points of the plurality of network elements.

14. The computing system of claim 11, wherein clustering the plurality of network elements into a plurality of clusters based on the determined one or more residual features of the plurality of network elements comprises clustering the plurality of network elements into a plurality of clusters without user equipment (UE) measurement reports (MRs), call history records (CHRs), configuration parameters, or engineering parameters of the plurality of network elements.

15. The computing system of claim 11, the computing system further configured to, prior to determining the global model, perform an additional layer of clustering by clustering the plurality of network elements into a plurality of super-clusters based on one or more features, and

wherein determining a global model comprises determining a global model for each of the plurality of super-clusters.

16. The computing system of claim 15, wherein the one or more features are determined based on the performance counter values and represent coverage, interference, or user equipment characteristics of the plurality of network elements.

17. The computing system of claim 15, wherein the one or more features comprises a correlation between an interference measurement and a traffic measurement of a network element or comprises a correlation between a call drop rate and an interference measurement of a network element.

18. The computing system of claim 11, the computing system further configured to, for each of the plurality of clusters, determine a respective regression model based on performance counter values and performance indicators of network elements within the cluster.

19. The computing system of claim 18, the computing system further configured to predict performance of a network element according to the regression model.

20. The computing system of claim 18, the computing system further configured to select cell physics features that are indicative of network performance behavior based on the plurality of clusters and the respective regression models.

21. A non-transitory, computer-readable medium storing computer-readable instructions executable by a computer and configured to perform operations comprising:

receiving a plurality of sets of data points of a plurality of network elements, each of the plurality of sets of data points corresponding to a respective network element of the plurality of network elements, the set of data points comprising performance counter values and a performance indicator of the respective network element;

determining a global model representing a global relationship pattern between the performance indicator and the performance counter values based on the plurality of sets of data points of the plurality of network elements;

for each network element of the plurality of network elements, determining one or more residual features, the one or more residual features based on error measures between the global model and the set of data points comprising the performance indicator and the performance counter values of the network element; and

clustering the plurality of network elements into a plurality of clusters based on the determined one or more residual features of the plurality of network elements.