Smart data cap avoidance with personalized predictions based on linear regression or historical usage alpha-generation patterns

Patent metadata

Field	Value
Publication number	US-10404541-B2
Application number	US-201715437118-A
Country	US
Kind code	B2
Filing date	Feb 20, 2017
Priority date	Jul 7, 2016
Publication date	Sep 3, 2019
Grant date	Sep 3, 2019

Abstract

A system, apparatus, and method for controlling data usage at a customer premises. A gateway configured as a single point of entry receives all data entering the premises. A DataCap Management Unit monitors all data usage by client devices within the premises and uses either a linear regression model or a historical usage alpha-generation method to analyze data usage patterns at the premises and to predict future data usage at the premises for a current billing cycle. The DataCap Management Unit dynamically adjusts a premises bandwidth cap throughout the current billing cycle, using the predicted future data usage as an input, to smoothly keep actual total data usage for the current billing cycle from exceeding a data cap before the current billing cycle ends.

First claim

What is claimed is:

1. A method implemented by a DataCap Management Unit for controlling data usage at a customer premises, the method comprising:
predicting future data usage by a plurality of client devices at the customer premises for a current billing cycle, wherein one or more of the client devices are engaged in adaptive video bitrate (ABR) stream flows, and wherein one or more of the client devices are engaged in non-ABR stream flows;
monitoring all actual data usage by the plurality of client devices at a single point of entry for the customer premises;
dynamically adjusting a premises bandwidth cap throughout the current billing cycle, using the predicted future data usage as an input, to keep actual total data usage for the current billing cycle from exceeding a data cap before the current billing cycle ends; and
throttling only the ABR stream flows and not the non-ABR stream flows to maintain a current premises bitrate within the premises bandwidth cap.

2. The method according to claim 1 , wherein predicting the future data usage includes generating a linear regression model for future data usage at the customer premises based on past data usage at the customer premises.

3. The method according to claim 2 , further comprising updating the linear regression model by recalculating coefficients as more recent data usage information becomes available.

4. The method according to claim 2 , wherein the linear regression model utilizes a least square method.

5. The method according to claim 2 , wherein the dynamically adjusting the premises bandwidth cap comprises:
applying the linear regression model to predict data consumption (Dc) for a current day;
applying the linear regression model to predict total data consumption (Dctotal) for remaining days of the current billing cycle;
calculating an offset of the Dc for an overrun or underrun (Dceffective) by adding to the Dc, a value: (data remaining within the data cap for the current billing cycle−Dctotal)/days remaining in the current billing cycle; and
calculating an hourly premises bandwidth cap by dividing the Dceffective by 24.

6. The method according to claim 5 , wherein there are a plurality of ABR stream flows active at the customer premises, and the method further comprises setting a bitrate limit for each ABR stream flow by dividing the hourly premises bandwidth cap by a number of the active ABR stream flows.

7. The method according to claim 2 , wherein the dynamically adjusting the premises bandwidth cap comprises:
applying the linear regression model to predict data consumption (Dc) for a current hour;
applying the linear regression model to predict total data consumption (Dctotal) for remaining hours of the current billing cycle; and
calculating an hourly premises bandwidth cap by adding to the Dc, a value: (data remaining within the data cap for the current billing cycle−Dctotal)/hours remaining in the current billing cycle.

8. The method according to claim 7 , wherein there are a plurality of ABR stream flows active at the customer premises, and the method further comprises setting a bitrate limit for each ABR stream flow by dividing the hourly premises bandwidth cap by a number of the active ABR stream flows.

9. The method according to claim 1 , wherein the predicting future data usage comprises generating a historical usage alpha-generation model for data usage at the customer premises based on historical data usage patterns at the customer premises during past billing cycles.

10. The method according to claim 9 , wherein the billing cycles are months, and when more than a year's worth of historical data usage information is available, generating a historical model for data usage at the customer premises includes generating a historical model for a current month utilizing data usage information from the same month from a previous year.

11. The method according to claim 9 , wherein the billing cycles are months, and when more than one month but less than a year's worth of historical data usage information is available, generating a historical model for data usage at the customer premises includes generating a historical model for a current month utilizing data usage information from a previous month.

12. The method according to claim 9 , wherein the throttling includes:
generating a value, α, equal to a time in the billing cycle (t m ) divided by a time spent consuming video (t c ) based on the historical data usage patterns at the customer premises during past billing cycles;
calculating, for a sum of all ABR stream flows entering the customer premises, an allowed data rate (ABR_Rate) by multiplying the α by a data remaining within the data cap divided by a number of days remaining in the billing cycle; and
setting a bitrate limit for each ABR stream flows based on the ABR_Rate and a number of ABR stream flows.

13. The method according to claim 1 , further comprising:
receiving an indication that a user desires to change a Datacap management type;
analyzing data flow;
generating a usage pattern graph based on a liner regression model;
generating a usage pattern graph based on a historical usage alpha-generation model;
displaying the usage pattern graphs for the user;
receiving an indication of a selection by the user to use either the linear regression model or the historical usage alpha-generation model to at least predict the future data usage;
applying datacap bandwidth management on data usage of a current month using either the linear regression model or the historical usage alpha-generation model which was selected by the user; and
applying a bandwidth limiter to a current calculation of allowed bandwidth using either the linear regression model or the historical usage alpha-generation model which was selected by the user.

14. The method according to claim 1 , further comprising:
receiving an indication that a user desires to have the DataCap Management unit to select to use either a linear regression model or a historical usage alpha-generation model to at least predict the future data usage; and
recognizing which of the linear regression model or the historical usage alpha-generation model is best to use based on examining changes over a time period and automatically switching to either the linear regression model or the historical usage alpha-generation model which offers a minimal change over the time period.

15. A DataCap Management Unit configured to control data usage at a customer premises, the DataCap Management Unit comprising:
an interface configured to receive all data entering the customer premises and to monitor all data usage by a plurality of client devices at the customer premises; and
a processing circuit configured to:
predict future data usage by the plurality of client devices at the customer premises for a current billing cycle, wherein one or more of the client devices are engaged in adaptive video bitrate (ABR) stream flows, and wherein one or more of the client devices are engaged in non-ABR stream flows;
dynamically adjust a premises bandwidth cap throughout the current billing cycle, using the predicted future data usage as an input, to keep actual total data usage for the current billing cycle from exceeding a data cap before the current billing cycle ends; and
throttle only the ABR stream flows and not the non-ABR stream flows to maintain a current premises bitrate within the premises bandwidth cap.

16. The DataCap Management Unit according to claim 15 , wherein the processing circuit is configured to predict the future data usage by generating a linear regres

Assignees

• Ericsson Telefon Ab L M
• Ericsson Ab

Inventors

• Phillips Chris
• Forsman Robert Hammond
• Reynolds Jennifer Ann
• Changrani Vishal

Classifications

• H04L41/0896 Primary

  Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities (flow or congestion control using dynamic resource allocation, e.g. in-call renegotiation, H04L47/76) · CPC title

• H04M15/8228

  Session based · CPC title

• H04M15/66

  Policy and charging system · CPC title

• H04L12/1407

  Policy-and-charging control [PCC] architecture · CPC title

• H04L12/1403

  Architecture for metering, charging or billing · CPC title