Server Resource Allocation Across Multiple Data Processing Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing resource allocation methods primarily optimize performance for a single data processing model, making it difficult to apply effectively in complex scenarios involving multiple data processing models.
Innovation Solution
A resource allocation method for a server with a multi-core architecture that involves obtaining tasks and data processing models, performing resource allocation on operators, and executing a coordinated resource allocation sub-method to allocate resources based on scheduling sequence and parallel execution state, while considering interference and utilization status.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing resource allocation methods are used to optimize performance for a single data processing model, then performance for that single model is improved, but the method becomes difficult to apply effectively in complex scenarios involving multiple data processing models
Solution Approach 1:
The resource allocation method is designed to handle both single-model and multi-model scenarios through a unified framework. The system can dynamically adapt to different numbers of data processing models (K=1 or K>1) and adjust its allocation strategy accordingly, making it universally applicable across various deployment scenarios without requiring separate optimization methods for each case.
Solution Approach 2:
The patent implements dynamic resource allocation that adapts to changing workloads and model configurations. The system continuously monitors the number of active data processing models and adjusts the resource allocation strategy in real-time, transitioning between single-model optimization and multi-model coordinated allocation based on current system state, thereby maintaining effectiveness across diverse scenarios.
2Reliability
If resources are allocated to multiple data processing models simultaneously, then service assurance and performance in complex scenarios are improved, but resource conflicts and interference between models may increase
Solution Approach 1:
The system performs preliminary resource allocation and scheduling before executing multiple data processing models. By pre-calculating resource requirements and establishing allocation plans for all K models in advance, the system prevents resource conflicts during execution. This proactive approach ensures that each operator receives its required resources without contention, thereby maintaining high service assurance while avoiding interference.
Solution Approach 2:
The patent introduces a coordinated resource allocation mechanism that acts as an intermediary between multiple data processing models. This mediator optimizes the scheduling sequence and parallel execution state of operators across different models, managing resource sharing and preventing interference by coordinating access to shared resources, thus enabling reliable multi-model execution.
3Loss of energy
If coordinated resource allocation is performed for multiple second tasks, then resource utilization is optimized and energy consumption is reduced, but the complexity of the allocation process increases
Solution Approach 1:
The coordinated resource allocation process is segmented into distinct phases: obtaining tasks and data processing models, performing initial resource allocation on operators, determining scheduling sequences, and executing parallel optimization. This segmentation breaks down the complex multi-model allocation problem into manageable sub-tasks, reducing implementation complexity while maintaining overall optimization effectiveness and energy efficiency.
Data Source
AI summary
A resource allocation method, a medium and a server are provided. The resource allocation method includes: obtaining tasks executable by the server as first tasks; obtaining first data processing models each corresponding to one of the first tasks, wherein each of the first data processing models includes one or more operators; performing a resource allocation on each operator in each of the first data processing models to obtain a quantity of resource used by the operator; and obtaining second tasks when the server receives a task request from a user, wherein the second tasks include current tasks of the server and tasks corresponding to the task request from the user; when the number of the second tasks is greater than one, a coordinated resource allocation sub-method is executed. The resource allocation method described in the present disclosure can be applied to complex scenarios involving multiple data processing models.