CN-122027816-A - Digital live calculation resource scheduling method, device, storage medium and computer program product

CN122027816ACN 122027816 ACN122027816 ACN 122027816ACN-122027816-A

Abstract

The application discloses a method, equipment, a storage medium and a computer program product for scheduling digital live broadcast computing resources, which relate to the technical field of computing resource scheduling and comprise the steps of acquiring a live broadcast preprocessing request sent by a user side; the method comprises the steps of calling a preset stream data model to generate a standardized data stream based on the live broadcast preprocessing request, triggering an asynchronous preprocessing task according to the standardized data stream, distributing the asynchronous preprocessing task to a resource preparation node in a preset computing power decoupling scheduling architecture according to a preset task priority distribution rule and a preset time window, and executing the asynchronous preprocessing task based on the resource preparation node and a differential computing logic. By adopting a stream data model, asynchronous task scheduling, a calculation force decoupling architecture and differential calculation, the calculation force utilization rate of digital live broadcasting is improved under the condition of guaranteeing the live broadcasting effect of the digital human.

Inventors

PENG ZHIQIANG
ZHU XIONG

Assignees

招商银行股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260313

Claims (10)

1. The digital live computing power resource scheduling method is characterized by comprising the following steps of: acquiring a live broadcast preprocessing request sent by a user; calling a preset stream data model to generate a standardized data stream based on the live broadcast preprocessing request, and triggering an asynchronous preprocessing task according to the standardized data stream; distributing the asynchronous preprocessing task to a resource preparation node in a preset computing power decoupling scheduling architecture according to a preset task priority distribution rule and a preset time window; the asynchronous preprocessing task is performed based on the resource preparation node and differential computing logic.
2. The method for scheduling digital live computing power resources according to claim 1, wherein the step of obtaining the live preprocessing request sent by the user terminal further comprises: Node separation is carried out on the resource preparation node and the real-time rendering node so as to construct mutually independent computing power clusters; and deploying connection components between mutually independent computing force clusters to obtain the computing force decoupling scheduling framework.
3. The method for scheduling digital human live computing power resources according to claim 1, wherein before the step of calling a preset stream data model to generate a standardized data stream based on the live preprocessing request, further comprises: carding discrete material types of the digital live scene, and respectively defining a unified metadata field set for the discrete material types; and establishing a stream data model based on the metadata field set, and reserving an increment fragment exclusive storage field in the stream data model.
4. The method for scheduling digital human live computing power resources according to claim 1, wherein the step of calling a preset stream data model based on the live preprocessing request to generate a standardized data stream and triggering an asynchronous preprocessing task according to the standardized data stream comprises: Analyzing the live broadcast preprocessing request, and pulling discrete materials based on an analysis result of the live broadcast preprocessing request; invoking the stream data model, inputting the discrete materials into the stream data model, and integrating and packaging according to a metadata field set defined by the stream data model to obtain a standardized data stream; and packaging the standardized data stream into a structured message and sending the structured message to a message queue to trigger an asynchronous preprocessing task.
5. The method of digital live computing power resource scheduling of claim 1, wherein the step of assigning the asynchronous preprocessing task to a resource preparation node in a preset computing power decoupling scheduling architecture according to a preset task priority assignment rule and a preset time window comprises: judging the task priority and the task time window of the asynchronous preprocessing task; if the task time window is in a rendering idle time window and the asynchronous preprocessing task is a low-priority preprocessing task, distributing the asynchronous preprocessing task to an idle rendering node; and if the task time window is not in the rendering idle time window, distributing the asynchronous preprocessing task to a resource preparation node according to the task priority.
6. The digital live computing power resource scheduling method of claim 1, wherein the step of performing the asynchronous preprocessing task based on the resource preparation node and differential computing logic comprises: Judging the task type of the asynchronous preprocessing task; if the task type is a full preprocessing task, loading a standardized data stream of the asynchronous preprocessing task, and generating a preprocessing resource packet based on the standardized data stream; If the task type is an incremental preprocessing task, differentially calculating an updated standardized data stream in the asynchronous preprocessing task, and generating a preprocessing resource packet based on the updated standardized data stream; performing an integrity check and a rendering compatibility check on the pre-processing resource package; And under the condition that the preprocessing resource package passes the integrity check and the rendering compatibility check, sending a resource ready signal to a rendering module so that the rendering module can call the preprocessing resource package to carry out digital live broadcasting.
7. The digital live computing power resource scheduling method of claim 1, wherein the step of performing the asynchronous preprocessing task based on the resource preparation node and differential computing logic further comprises: monitoring a live broadcast temporary modification request sent by a user; Triggering a temporary preprocessing task based on the live temporary modification request, and distributing the temporary preprocessing task to a standby resource preparation node so as to execute the temporary preprocessing task.
8. A digital live computing power resource scheduling device, characterized in that the device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, the computer program being configured to implement the steps of the digital live computing power resource scheduling method according to any one of claims 1 to 7.
9. A storage medium, characterized in that the storage medium is a computer readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements the steps of the digital live computing power resource scheduling method according to any one of claims 1 to 7.
10. A computer program product, characterized in that the computer program product comprises a computer program which, when executed by a processor, implements the steps of the digital live calculation resource scheduling method of any one of claims 1 to 7.

Description

Digital live calculation resource scheduling method, device, storage medium and computer program product Technical Field The present application relates to the technical field of computing power resource scheduling, and in particular, to a digital live computing power resource scheduling method, a digital live computing power resource scheduling device, a digital live computing power resource scheduling storage medium, and a digital live computing power resource scheduling computer program product. Background In the industry, digital live broadcasting generally adopts a preprocessing and rendering coupling deployed computing power architecture, all computing tasks are executed on the same rendering node cluster, and exclusive scheduling rules are not designed aiming at the business characteristics of digital live broadcasting. The rendering node only bears high-load tasks in the live broadcast on-demand period, the CPU utilization rate of the rendering node in the non-on-demand period is generally low, a large amount of computing power is idle, meanwhile, the preprocessing task has no exclusive computing power resource, the computing power needs to be queued in the peak period, the overall computing power utilization rate of the whole cluster is insufficient, and the computing power cost is high. And as the preprocessing task and the rendering task have no separation in calculation, the live broadcasting stability is difficult to ensure, and the incremental modification response efficiency proposed by a user is low. Therefore, how to improve the calculation power utilization rate of digital live broadcasting under the condition of guaranteeing the live broadcasting effect of the digital live broadcasting becomes a technical problem to be solved in the application. Disclosure of Invention The application mainly aims to provide a digital live-broadcasting calculation resource scheduling method, equipment, a storage medium and a computer program product, which aim to solve the technical problem of improving the calculation utilization rate of live broadcasting of a digital person under the condition of guaranteeing the live broadcasting effect of the digital person. In order to achieve the above purpose, the application provides a digital live calculation power resource scheduling method, which comprises the following steps: acquiring a live broadcast preprocessing request sent by a user; calling a preset stream data model to generate a standardized data stream based on the live broadcast preprocessing request, and triggering an asynchronous preprocessing task according to the standardized data stream; distributing the asynchronous preprocessing task to a resource preparation node in a preset computing power decoupling scheduling architecture according to a preset task priority distribution rule and a preset time window; the asynchronous preprocessing task is performed based on the resource preparation node and differential computing logic. In an embodiment, before the step of obtaining the live broadcast preprocessing request sent by the user side, the method further includes: Node separation is carried out on the resource preparation node and the real-time rendering node so as to construct mutually independent computing power clusters; and deploying connection components between mutually independent computing force clusters to obtain the computing force decoupling scheduling framework. In an embodiment, before the step of calling a preset stream data model to generate a standardized data stream based on the live broadcast preprocessing request, the method further includes: carding discrete material types of the digital live scene, and respectively defining a unified metadata field set for the discrete material types; and establishing a stream data model based on the metadata field set, and reserving an increment fragment exclusive storage field in the stream data model. In an embodiment, the step of calling a preset stream data model to generate a standardized data stream based on the live broadcast preprocessing request and triggering an asynchronous preprocessing task according to the standardized data stream includes: Analyzing the live broadcast preprocessing request, and pulling discrete materials based on an analysis result of the live broadcast preprocessing request; invoking the stream data model, inputting the discrete materials into the stream data model, and integrating and packaging according to a metadata field set defined by the stream data model to obtain a standardized data stream; and packaging the standardized data stream into a structured message and sending the structured message to a message queue to trigger an asynchronous preprocessing task. In an embodiment, the step of allocating the asynchronous preprocessing task to the resource preparation node in the preset computing power decoupling schedule architecture according to a preset task priority allocation rule and a preset time windo