Delay-constrained multi-path computing power routing method in computing power network

By constructing a virtual network model and solving convex problems to optimize data allocation, the problem of comprehensively selecting computing resources and network paths in computing power routing is solved, realizing low-cost and low-latency multi-path routing in computing power networks, and meeting the needs of users' latency-sensitive computing tasks.

CN117880165BActive Publication Date: 2026-06-09ZHEJIANG NEW INTERNET EXCHANGE CENT CO LTD +1
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
ZHEJIANG NEW INTERNET EXCHANGE CENT CO LTD
Filing Date
2023-12-23
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing computing power routing technologies fail to effectively combine computing resources and network paths for comprehensive selection, resulting in high network transmission and computing costs and difficulty in meeting the needs of users' latency-sensitive computing tasks.

Method used

A multi-path computing power routing algorithm is designed. By constructing a virtual network model and combining computing power resources and network path latency constraints, the algorithm optimizes data volume allocation to minimize total cost and meet maximum latency constraints. Convex problem solving and traffic decomposition techniques are used for path selection.

Benefits of technology

It achieves the minimization of the total cost of network transmission and computation in user latency-sensitive computing tasks, and obtains near-optimal performance in a shorter time, reducing latency constraints and user task data volume violation rates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117880165B_ABST
    Figure CN117880165B_ABST
Patent Text Reader

Abstract

The application discloses a kind of delay constraint multi-path computing power routing methods in computing power network, comprising the following steps: constructing the virtual network model of computing power network, by increasing virtual node as terminal, computing power is mapped as the link capability between computing power node and virtual node.The maximum delay constraint of multi-path computing power routing problem in virtual network model is approximated into total delay constraint, and the approximate optimization result under total delay constraint is obtained.Under the condition of guaranteeing maximum delay constraint, data volume transfer is carried out between maximum unit data volume cost path and minimum unit data volume cost path.The multi-path computing power routing algorithm proposed can solve the common selection problem of computing power resources for computing and network path for transmission, obtain approximate optimal total cost and meet maximum delay limit.The application can be used for multi-path computing power routing of delay-sensitive computing tasks in computing power network.
Need to check novelty before this filing date? Find Prior Art