VNFM Distributed Resource Allocation Bypassing NFVO

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In network functions virtualization (NFV) systems, the reliance on a single point of failure in NFVO (NFV Orchestrator) for resource allocation and management leads to significant network impacts when faults occur, compromising robustness and efficiency.

Innovation Solution

Implementing a distributed resource management method where VNFM (Virtualized Network Function Manager) can request resources directly from peer VNFM without relying on NFVO, forming a resource management network between VNFM to enable resource allocation and query processes, thereby bypassing NFVO participation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If NFVO is used for resource allocation and management, then centralized control is achieved, but network robustness deteriorates due to single point of failure

Engineering Contradiction:
Improvecentralized controlVSAvoidnetwork robustness
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent segments the centralized NFVO resource management function into distributed VNFM nodes. Each VNFM can independently handle resource allocation requests from other VNFMs, transforming the single-point centralized control into a distributed multi-node architecture that eliminates the single point of failure while maintaining operational control capabilities

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces VNFM as an intermediary node between VNFs and NFVO. VNFMs can directly communicate with each other for resource allocation without requiring NFVO intervention, acting as mediators that enable peer-to-peer resource transactions while reducing dependency on the centralized NFVO

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If NFVO is used for resource allocation, then centralized management is achieved, but network efficiency deteriorates when NFVO encounters faults

Engineering Contradiction:
Improvecentralized managementVSAvoidnetwork efficiency
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent divides the centralized resource management workload into multiple distributed VNFM nodes that can independently process resource allocation requests. This segmentation allows parallel processing of resource requests across multiple VNFMs simultaneously, preventing the bottleneck that occurs when all requests must pass through a single NFVO, thereby maintaining high network efficiency even as system scale increases

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent enables VNFMs to autonomously handle resource allocation by allowing them to directly query and allocate resources from VIMs connected to peer VNFMs without requiring NFVO mediation. This self-service capability eliminates the efficiency bottleneck of centralized sequential processing, allowing each VNFM to independently and simultaneously manage its own resource allocation operations

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP3358790B1Network function virtualization resource processing method and virtualized network function manager
Publication Date: 2020.03.18 HUAWEI TECH CO LTD
  • EP3358790B1 patent drawingFigure 1
  • EP3358790B1 patent drawingFigure 2
  • EP3358790B1 patent drawingFigure 3~4

AI summary

The present invention provides a network functions virtualization resource processing method and a virtualized network function manager. A first VNFM sends, without using an NFVO, a first resource request message for requesting a resource to a second VNFM connected to the first VNFM. When a resource of a target VIM connected to the second VNFM meets information about a resource that is applied for, the first VNFM determines, according to a received first feedback message, information about the target VIM that can reserve or allocate the resource that is applied for. A resource management network is formed between the first VNFM and the second VNFM, so that the interconnected VNFMs can perform resource processing such as resource query or resource allocation between each other by using an interface. This process does not require participation of the NFVO. Therefore, impact on the network if the NFVO is faulty is avoided, and a distributed resource management manner is implemented. In this way, when the NFVO is faulty, the first VNFM can still manage the resource that is applied for, effectively improving robustness and efficiency of the network.