Methods and systems for profiling network resource usage by a mobile application

Patent metadata

Field	Value
Publication number	US-9693355-B2
Application number	US-201514805361-A
Country	US
Kind code	B2
Filing date	Jul 21, 2015
Priority date	Jul 21, 2015
Publication date	Jun 27, 2017
Grant date	Jun 27, 2017

Abstract

An exemplary profiling system builds a two-layer mapping model for a mobile network. The two-layer mapping model establishes a causal relationship between a plurality of application behavior indicators and network resource usage within the mobile network by defining a first mapping relationship between the plurality of application behavior indicators and a plurality of network performance indicators representative of network traffic that passes through the mobile network, and a second mapping relationship between the plurality of network performance indicators and network resource usage within the mobile network. Corresponding systems and methods are also described.

First claim

What is claimed is:

1. A method comprising:
building, by a profiling system, a two-layer mapping model for use in profiling network resource usage by a mobile application executed by one or more mobile devices that operate on a mobile network, the building comprising
acquiring, from a plurality of mobile devices, application behavior data collected by the mobile devices during a training time period, the application behavior data including values for a plurality of application behavior indicators associated with a plurality of mobile applications executed by the mobile devices while the mobile devices operate within a test cell included within the mobile network;
acquiring network traffic data and network resource usage data both corresponding to the test cell, wherein the network traffic data includes values for a plurality of network traffic indicators representative of network traffic that passes through the test cell during the training time period and wherein the network resource usage data represents an amount of network resource usage at the test cell during the training time period;
identifying a subset of network traffic indicators included in the plurality of network traffic indicators and that are relatively more correlated to the network resource usage at the test cell during the training time period than a remaining number of network traffic indicators included in the plurality of network traffic indicators;
identifying a subset of application behavior indicators included in the plurality of application behavior indicators and that are relatively more correlated to the subset of network traffic indicators than a remaining number of application behavior indicators included in the plurality of application behavior indicators; and
determining a causal relationship between the subset of application behavior indicators and network resource usage at any cell within the mobile network by
determining a first mapping relationship between the subset of application behavior indicators and the subset of network traffic indicators by regressing the subset of application behavior indicators to the subset of network traffic indicators, and
determining a second mapping relationship between the subset of network traffic indicators and the network resource usage at any cell within the mobile network by regressing the subset of network traffic indicators to the network resource usage data; and

using, by the profiling system and subsequent to the building of the two-layer mapping model, the two-layer mapping model to facilitate management of one or more mobile network resources by
accessing application behavior data specific to a particular mobile application included in the plurality of mobile applications; and
profiling network resource usage by the particular mobile application within the mobile network by applying the application behavior data specific to the particular mobile application to the two-layer mapping model.

2. The method of claim 1 , wherein the identifying of the subset of network traffic indicators included in the plurality of network traffic indicators and that are relatively more correlated to the network resource usage at the test cell during the training time period than the remaining number of network traffic indicators included in the plurality of network traffic indicators comprises:
generating a set of importance scores for the plurality of network traffic indicators, the set of importance scores indicating a degree of correlation of each network traffic indicator included in the plurality of network traffic indicators to the network resource usage data;
selecting, based on the set of importance scores, various network traffic indicators included in the plurality of network traffic indicators for inclusion in the subset of network traffic indicators.

3. The method of claim 2 , wherein the generating and selecting are performed in accordance with a proximity matrix-assisted feature selection heuristic.

4. The method of claim 2 , wherein the selecting of the various network traffic indicators for inclusion in the subset of network traffic indicators comprises selecting a particular network traffic indicator included in the plurality of network traffic indicators for inclusion in the subset of network traffic indicators if the particular network traffic indicator has an importance score above a predetermined threshold.

5. The method of claim 2 , wherein the selecting of the various network traffic indicators for inclusion in the subset of network traffic indicators comprises selecting, for inclusion in the subset of network traffic indicators, a predetermined number of network traffic indicators included in the plurality of network traffic indicators and that have higher importance scores than the remaining number of network traffic indicators.

6. The method of claim 1 , wherein the identifying of the subset of application behavior indicators included in the plurality of application behavior indicators and that are relatively more correlated to the subset of network traffic indicators than the remaining number of application behavior indicators included in the plurality of application behavior indicators comprises:
generating a set of importance scores for the plurality of application behavior indicators, the set of importance scores indicating a degree of correlation of each application behavior indicator included in the plurality of application behavior indicators to the subset of network traffic indicators;
selecting, based on the set of importance scores, various application behavior indicators included in the plurality of application behavior indicators for inclusion in the subset application behavior indicators.

7. The method of claim 6 , wherein the generating and selecting are performed in accordance with a proximity matrix-assisted feature selection heuristic.

8. The method of claim 6 , wherein the selecting of the various application behavior indicators for inclusion in the subset of application behavior indicators comprises selecting a particular application behavior indicator included in the plurality of application behavior indicators for inclusion in the subset of application behavior indicators if the particular application behavior indicator has an importance score above a predetermined threshold.

9. The method of claim 6 , wherein the selecting of the various application behavior indicators for inclusion in the subset of application behavior indicators comprises selecting, for inclusion in the subset of application behavior indicators, a predetermined number of application behavior indicators included in the plurality of application behavior indicators and that have higher importance scores than the remaining number of application behavior indicators.

10. The method of claim 1 , wherein the regressing of the subset of application behavior indicators to the subset of network traffic indicators and the regressing of the subset of network traffic indicators to the network resource usage data are performed in accordance with a sliding-window-based local weight scatterplot smoothing heuristic.

11. The method of claim 1 , embodied as computer-executable instructions on at least one non-transitory computer-readable medium.

12. A method comprising:
acquiring, by a profiling system, application behavior data specific to a particular mobile application executed by one or more mobile devices while the one or more mobile devices operate on a mobile network, the application behavior data including values for a subset of application behavior indicators included in a plurality of application behavior indicators associated with the particular mobile appli

Assignees

• Verizon Patent & Licensing Inc

Inventors

• Ouyang Ye
• Becht Carol

Classifications

• H04L43/08

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

• H04W24/10

  Scheduling measurement reports {; Arrangements for measurement reports} · CPC title

• H04W72/048 Primary

  Electricity · mapped topic

• H04W24/08 Primary

  Testing, {supervising or monitoring} using real traffic · CPC title

• H04W72/51

  based on terminal or device properties · CPC title