Method for dynamic computational resource management and apparatus for implementing the same

Patent metadata

Field	Value
Publication number	US-12058397-B2
Application number	US-202318098443-A
Country	US
Kind code	B2
Filing date	Jan 18, 2023
Priority date	Jan 18, 2022
Publication date	Aug 6, 2024
Grant date	Aug 6, 2024

Abstract

A method for managing computational resources allocated for encoding of one or more multimedia content streams for distribution in dynamic mode to viewing devices through a distribution network is proposed, which comprises, by a processing node of the distribution network: obtain, for a multimedia content stream corresponding to a channel distributed to the viewing devices, a value of a computational resource allocation criterion, wherein the computational resource allocation criterion comprises an audience measurement for the corresponding channel and a video content complexity of the multimedia content stream; and determine, based on the computational resource allocation criterion, an allocation of computational resources of a computing platform configured for encoding the multimedia content stream.

First claim

The invention claimed is:

1. A method for managing computational resources allocated for encoding of one or more multimedia content streams for distribution in dynamic mode to viewing devices through a distribution network, the method comprising, by a processing node of the distribution network:
obtaining, for a multimedia content stream corresponding to a channel distributed to the viewing devices, a value of a computational resource allocation criterion, wherein the computational resource allocation criterion comprises an audience measurement for the corresponding channel and a video content complexity of the multimedia content stream;
determining, based on the computational resource allocation criterion, an allocation of computational resources of a computing platform configured for encoding the multimedia content stream;
obtaining respective computational resource allocation criterion levels for a plurality of multimedia content streams respectively corresponding to channels in a plurality of channels distributed to viewers;
performing one or more iterations of a channel allocation loop, wherein an iteration of index i comprises, for a current channel of the plurality of channels: determine for the current channel a respective channel gain based on a rate-distortion criterion J i =R i +λ·D i , where D i is a distortion measurement, λ is a Lagrangian parameter, and R i is an overall bitrate measurement, wherein R i is determined based on an audience measurement for the current channel and an estimate of gain in bitrate reduction obtained if the current channel is allocated additional computation resources for encoding the corresponding multimedia content stream, and wherein D i is determined based on a distortion measurement and an audience measurement for the current channel;
selecting, in the plurality of channels, a channel based on one or more respective channel gains; and
allocating additional computational resources to encoding the multimedia content stream corresponding to the selected channel.

2. The method according to claim 1 , further comprising:
determining an allocation of computational resources that minimizes an overall bitrate of a plurality of multimedia content streams after encoding, while maximizing a video quality by minimizing a distortion incurred by encoding the plurality of multimedia content streams.

3. The method according to claim 1 , further comprising:
determining an allocation of computational resources that minimizes an overall bitrate of a plurality of multimedia content streams after encoding, while maximizing a quality of experience for channels corresponding to the plurality of multimedia content streams.

4. The method according to claim 1 , further comprising:
determining, for the channel, an estimate of gain in bitrate reduction obtained if the channel is allocated additional computation resources for encoding the corresponding multimedia content stream.

5. The method according to claim 1 , wherein the overall bitrate measurement R i is determined based on a product of v i with the estimate of gain, wherein v i is an audience measurement for the current channel.

6. The method according to claim 1 , wherein the overall bitrate measurement R t is determined based on a product of (1+v i ) with the estimate of gain, wherein v i is an audience measurement for the current channel.

7. The method according to claim 1 , wherein the distribution network comprises a content delivery network, CDN, wherein the overall bitrate measurement R i is determined based on a product of (Σ j=1 M k j +v i ) with the estimate of gain, wherein v i is an audience measurement for the current channel, M is a number of regions in the CDN, and k j is a weight value of index j assigned to an edge cache used for distribution of the current channel in the CDN.

8. The method according to claim 1 , wherein the distortion measurement of D i is determined based on d i , a distortion measurement for the current channel, and/or QoE i , a Quality of Experience measurement for the current channel.

9. The method according to claim 8 , wherein the distortion measurement of D i is determined based on the Quality of Experience measurement for the current channel QoE i , and wherein the Quality of Experience, QoE, measurement for the current channel is determined based on one or more of the following QoE metrics: average of time taken by each player to start playback; average of rebuffering time for all viewers; average of latencies to get the frame for all viewers; number of failed playback attempts; average number of dropped frames for all viewers of the channel; and average of rebuffering frequency for all viewers of the channel.

10. The method according to claim 4 , wherein the estimate of gain in bitrate reduction is determined based on a prediction of bitrate gain determined using a machine learning algorithm.

11. The method according to claim 10 , wherein the machine learning algorithm is configured to receive input data corresponding to one or more of the following input data parameters: current CPU allocation, current video quality, estimated video content complexity, video content frame width, video content frame height, video content frame rate, interlaced or progressive video content, and type of encoder.

12. The method according to claim 10 , wherein the machine learning algorithm is a supervised learning algorithm, the method further comprising: performing a training phase for training the supervised learning algorithm during which the supervised learning algorithm is trained with training data.

13. The method according to claim 12 , wherein the supervised learning algorithm is configured to use online learning for performing an ongoing training during an inference phase of the supervised learning algorithm.

14. The method according to claim 4 , wherein the additional computational resources comprise an additional CPU core used for encoding the corresponding multimedia content stream.

15. An apparatus, the apparatus comprising a processor and a memory operatively coupled to the processor, wherein the apparatus is configured to perform a method for managing computational resources allocated for encoding of one or more multimedia content streams for distribution in dynamic mode to viewing devices through a distribution network, the method comprising, by a processing node of the distribution network:
obtaining, for a multimedia content stream corresponding to a channel distributed to the viewing devices, a value of a computational resource allocation criterion, wherein the computational resource allocation criterion comprises an audience measurement for the corresponding channel and a video content complexity of the multimedia content stream; and
determining, based on the computational resource allocation criterion, an allocation of computational resources of a computing platform configured for encoding the multimedia content stream;
obtaining respective computational resource allocation criterion levels for a plurality of multimedia content streams respectively corresponding to channels in a plurality of channels distributed to viewers;
performing one or more iterations of a channel allocation loop, wherein an iteration of index i comprises, for a current channel of the plurality of channels: determine for the current channel a respective channel gain based on a rate-distortion criterion J i =R i +λ·D i , where D i is a distortion measurement, λ is a Lagrangian parameter, and R i is an overall bitrate measurement, wherein R i is determined based on an audience measurement for the current channe

Assignees

• Ateme

Inventors

• Moussaoui Abdelmajid
• Guionnet Thomas
• Pejhan Sassan

Classifications

• H04N21/2405

  Monitoring of the internal components or processes of the server, e.g. server load (allocation of resources in multiprogramming arrangements G06F9/50; performance measurement of computer activity G06F11/34) · CPC title

• H04N21/2402

  Monitoring of the downstream path of the transmission network, e.g. bandwidth available (traffic monitoring in data switching networks H04L43/00; monitoring data switching networks utilization H04L43/0876) · CPC title

• H04N21/2385

  Channel allocation (H04N21/266 takes precedence); Bandwidth allocation (H04N21/24 takes precedence {; allocation of channels according to the instantaneous demands of the users in time-division multiplex systems H04J3/1682; admission control, resource allocation in open networks H04L12/5692; arrangements for maintenance or administration in data switching networks involving bandwidth and capacity management H04L41/0896; negotiating bandwidth in wireless networks H04W28/16}) · CPC title

• H04N19/164

  Feedback from the receiver or from the transmission channel · CPC title

• H04N19/156

  Availability of hardware or computational resources, e.g. encoding based on power-saving criteria · CPC title