Network card distribution method and device and electronic equipment

By dividing network card sets according to the affinity relationship between network card and service in the NUMA architecture server and setting priority, the problem of failing to fully utilize the advantages of NUMA architecture in the traditional network card allocation method is solved, and more efficient network communication and resource scheduling is achieved.

CN119728432AActive Publication Date: 2025-03-28INSPUR SUZHOU INTELLIGENT TECH CO LTD
5 Cites 0 Cited by

Patent Information

Application Number
CN202412000435.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-31
Publication Date
2025-03-28
Estimated Expiration
2044-12-31

AI Technical Summary

Technical Problem

In servers with non-unified memory access (NUMA) architecture, traditional network card allocation methods fail to fully consider the multi-level affinity relationship between network cards and services, resulting in improper allocation of network services, increasing memory access latency, and reducing system network communication efficiency.

Method used

By determining whether the target network card has affinity with the access node and processor slot where the target service is located, the network card is divided into multiple sets, and prioritization is set according to the affinity relationship, and network cards with a high affinity with the target service are preferred for allocation.

Benefits of technology

Reduces memory access latency across NUMA nodes or slots, improves network communication efficiency, achieves more refined resource scheduling, and avoids resource waste and performance bottlenecks.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a network card distribution method and device and electronic equipment, and relates to the technical field of information, and the method comprises the steps: judging whether a target network card is compatible with an access node and / or a processor slot where a target service is located, and determining whether the target network card is compatible with the access node and / or the processor slot based on a judgment result; the target network card is correspondingly updated to a first network card set, a second network card set or a third network card set, the second network card set comprises network cards compatible with a processor slot where the target service is located, and the third network card set comprises network cards compatible with an access node where the target service is located; and selecting the network card with the highest priority from the first network card set, the second network card set and the third network card set as the network card to be allocated for the target service. The optimal matching between the service and the network card is ensured through multi-level sorting and accurate network card selection, and the service running efficiency is improved especially in the complex environment of multiple NUMA nodes and multiple processor slots.
Need to check novelty before this filing date? Find Prior Art