- Number
- 20260267708
- Published
- 2026-09-10
- Filed
- 2026-04-24
- Assignee
- Disney Enterprises, Inc.
- Inventors
- Howe; Thomas, Kegel; Mark, Russell; Edward, Higgins; Joseph, Spiewak; Daniel, Sainz; Matthias
- CPC
- G06F9/505
- Verdict
- Set aside cdn/infra
- In edition
- 2026-W37
- Source
- Google Patents · FreePatentsOnline
The keeper's note
a system receives usage data for content delivered by multiple content delivery network (CDN) entities during a last time interval.
Abstract
In some embodiments, a system receives usage data for content delivered by multiple content delivery network (CDN) entities during a last time interval. This usage data is based on delivering the content from servicing first requests. The system calculates ideal distribution weights from this usage data to allocate second requests to the CDN entities in a next time interval. It also calculates effective distribution weights from the usage data to allocate the second requests, where these effective weights are further based on a current network condition. A selection parameter is then generated based on correcting the effective distribution weights to be closer to the ideal distribution weights for the next time interval. The system outputs this selection parameter and uses it to select a particular CDN entity for a second request. This selection is performed according to a distribution for the next time interval that is closer to the ideal distribution than it is to the effective distribution.
Background
BACKGROUND
A content provider may use multiple content delivery network entities to deliver content to client devices. For example, content delivery network entities may deliver videos for a video service provider. A content delivery network entity may be a cluster of servers that can deliver content for the provider. When a client device requests a video, a service of the provider may select one of the content delivery network entities to process the request. Upon receiving a selection of a content delivery network entity, the client device communicates with that content delivery network entity to have the video delivered from that content delivery network entity. Each content delivery network entity may have an associated policy that provides parameters that are used by the service to select a content delivery network entity for a request. The parameters are static unless they are changed manually. However, the conditions of how videos are delivered during different time periods may rapidly change, such as at the top of the hour when new shows become available on live television or when a show becomes popular. The use of static parameters may not optimally select content delivery network entities when the conditions change rapidly.
Claims
1. A method executed by a computing device comprising: receiving usage data for content that was delivered in a last time interval by a plurality of content delivery network (CDN) entities, wherein: the usage data is based on delivering the content from servicing first requests, ideal distribution weights are based on the usage data to allocate second requests to the plurality of CDN entities in a next time interval, effective distribution weights based on the usage data to allocate the second requests to the plurality of CDN entities in the next time interval, the effective distribution weights being based a network condition, and a selection parameter is based on a correction of the effective distribution weights to be closer to the ideal distribution weights in the next time interval, outputting the selection parameter; and selecting, based on the selection parameter, a particular CDN entity of the plurality of CDN entities for a second request, wherein the selecting is performed according to a distribution that is closer to an ideal distribution than an effective distribution in the next time interval. ||
19. A non-transitory computer-readable storage medium containing instructions, that when executed, control a computer system to be operable for: receiving usage data for content that was delivered in a last time interval by a plurality of content delivery network (CDN) entities, wherein: the usage data is based on delivering the content from servicing first requests, ideal distribution weights are based on the usage data to allocate second requests to the plurality of CDN entities in a next time interval, effective distribution weights based on the usage data to allocate the second requests to the plurality of CDN entities in the next time interval, the effective distribution weights being based a network condition, and a selection parameter is based on a correction of the effective distribution weights to be closer to the ideal distribution weights in the next time interval, outputting the selection parameter; and selecting, based on the selection parameter, a particular CDN entity of the plurality of CDN entities for a second request, wherein the selecting is performed according to a distribution that is closer to an ideal distribution than an effective distribution in the next time interval. ||
20. An apparatus comprising: one or more computer processors; and a non-transitory computer-readable storage medium comprising instructions, that when executed, control the one or more computer processors to be operable for: receiving usage data for content that was delivered in a last time interval by a plurality of content delivery network (CDN) entities, wherein: the usage data is based on delivering the content from servicing first requests, ideal distribution weights are based on the usage data to allocate second requests to the plurality of CDN entities in a next time interval, effective distribution weights based on the usage data to allocate the second requests to the plurality of CDN entities in the next time interval, the effective distribution weights being based a network condition, and a selection parameter is based on a correction of the effective distribution weights to be closer to the ideal distribution weights in the next time interval, outputting the selection parameter; and selecting, based on the selection parameter, a particular CDN entity of the plurality of CDN entities for a second request, wherein the selecting is performed according to a distribution that is closer to an ideal distribution than an effective distribution in the next time interval.