Compression configuration identification

Patent metadata

Field	Value
Publication number	US-9681332-B2
Application number	US-201414316653-A
Country	US
Kind code	B2
Filing date	Jun 26, 2014
Priority date	Jun 26, 2014
Publication date	Jun 13, 2017
Grant date	Jun 13, 2017

Abstract

Apparatuses, methods and storage media associated with file compression and transmission, or file reception and decompression. Specifically, one or more compression/decompression or transmission/reception parameters associated with transmission or reception may be identified. Based on the identified parameters, energy consumption of compression and transmission, or reception and decompression, of the data over a wireless communication link may be predicted. Based on that prediction, a compression configuration may be identified. Other embodiments may be described and/or claimed.

First claim

What is claimed is:

1. A computer apparatus comprising:
a transceiver communication module comprising first computer program instructions to transmit or receive a compressed data over a communication link; and
a compression configuration module coupled with the transceiver communication module, the compression configuration module comprising second computer program instructions to:
identify one or more compression/decompression parameters, a length of time and energy of compression/decompression associated with compression of a data to form the compressed data or decompression of the compressed data to form the data, and a transmission/reception energy associated with transmission or reception of the compressed data over the communication link;
predict with a trained machine learning algorithm a predicted energy consumption based on an energy consumption trade-off among the identified one or more compression/decompression parameters, the length of time and energy of compression/decompression, and the transmission/reception energy, for each of a plurality of compression configurations;
identify, based on the predicted energy consumption, a compression configuration from the plurality of compression configurations;
obtain an energy and time outcome of the compression configuration; and
wherein a difference between the predicted energy consumption and the obtained energy and time outcome of the identified compression configuration is used to retrain the trained machine learning algorithm.

2. The apparatus of claim 1 , wherein the compression configuration in the plurality of compression configurations includes an indication of a compression utility, an indication of a compression level, or an indication of a central processing unit (CPU) frequency and wherein the trained machine learning algorithm is a random forest algorithm.

3. The apparatus of claim 1 , wherein the compression/decompression parameters comprise compression applied, file type of the data, and file size of the data.

4. The apparatus of claim 1 , wherein the transmission/reception energy is based on a prediction of a time requirement for the transmission of the compressed data.

5. The apparatus of claim 1 , wherein the data are data to be transmitted to an application server of a network, and wherein the transceiver communication module is to further:
compress the data to form the compressed data using the identified compression configuration; and
then transmit the compressed data to the application server over the communication link.

6. The apparatus of claim 1 , wherein the compressed data are compressed data to be received from an application server of a network, and wherein the transceiver communication module is to further transmit an indication of the compression configuration to the application server to use to compress the data before sending the compressed data to the apparatus over the communication link.

7. One or more non-transitory computer-readable media comprising instructions that, when executed by one or more processors of a mobile device, cause the mobile device to:
identify a data to be transferred or received over a communication link;
predict with a trained machine learning algorithm a predicted energy consumption based on an energy consumption trade-off among a transmit/receive energy of a compressed data in the communication link and compression/decompression parameters that include an energy consumption of compression/decompression of the data/the compressed data for each of a plurality of compression configurations;
identify, based on the predicted energy consumption, a compression configuration;
compress the data and transmit the compressed data, or receive and decompress the compressed data using the identified compression configuration;
obtain an energy and time outcome of the identified compression configuration;
wherein a difference between the predicted energy consumption and the obtained energy and time outcome of the identified compression configuration is used to retrain the trained machine learning algorithm.

8. The one or more computer-readable media of claim 7 , wherein the compression configuration in the plurality of compression configurations includes an indication of a compression utility, an indication of a compression level, or an indication of a central processing unit (CPU) frequency.

9. The one or more computer-readable media of claim 7 , wherein the predicted energy consumption is further based on available bandwidth of the communication link, condition of the communication link or communication technology, and wherein the compression/decompression parameters include compression applied, file type of the data, file size of the data, and a compression/decompression time.

10. The one or more computer-readable media of claim 7 , wherein the trained machine learning algorithm is trained with information about a previous compression configuration for a previously transmitted data and a difference between a previously predicted energy consumption for the previously transmitted data and a previous energy and time outcome of the previous compression configuration.

11. The one or more computer-readable media of claim 10 , wherein the trained machine learning algorithm is a random forest machine learning algorithm.

12. The one or more computer-readable media of claim 7 , wherein the predicted energy consumption is based at least in part on a prediction of a time requirement for compression and transmission or reception and decompression.

13. The one or more computer-readable media of claim 7 , wherein the communication link is a wireless communication link.

14. A method comprising:
predicting by a mobile device with a trained machine learning algorithm a predicted energy consumption wherein the predicted energy consumption is based on, for each of a plurality of compression configurations, an energy consumption trade-off among a time and energy required to compress a data into a compressed data or a time and energy to decompress the compressed data into the data, compression/decompression parameters, and an energy of transmission of the compressed data over a communication link or an energy of reception of the compressed data over the communication link, wherein the data is to be compressed to form the compressed data using a compression configuration of the plurality of compression configurations on the data;
identifying, by the mobile device, based on the predicted energy consumption, the compression configuration to be used to reduce energy consumption;
facilitating, by the mobile device, compression or decompression of the data using the identified compression configuration;
obtaining, by the mobile device, an energy and time outcome of the identified compression configuration;
wherein a difference between the predicted energy consumption and the energy and time outcome of the compression configuration is used to retrain the trained machine learning algorithm.

15. The method of claim 14 , wherein a compression configuration in the plurality of compression configurations includes an indication of a compression utility, an indication of a compression level, or an indication of a central processing unit (CPU) frequency.

16. The method of claim 14 , wherein the predicted energy consumption is further based on available bandwidth of the communication link, condition of the communication link or communication technology, and compression/decompression parameters that include compression applied, file type of the data, or file size of the data.

17. The method of claim

Assignees

• Intel Corp

Inventors

• Min Alexander W
• Lin Guan-Yu
• Tai Tsung-Yuan C
• Tsai Jr-Shian James

Classifications

• H04L12/12

  Arrangements for remote connection or disconnection of substations or of equipment thereof · CPC title

• H04W52/0251

  using monitoring of local events, e.g. events related to user activity · CPC title

• H04L45/32

  Flooding (denial of service attacks H04L63/1458) · CPC title

• H04L45/26

  Route discovery packet · CPC title

• H04L43/08

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