Virtualized Network Function Placement Balancing Latency and Server Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing VNF placement systems in telecommunications networks often assume fixed packet delays and impose constraints on latency, failing to efficiently manage traffic and resource utilization across heterogeneous server types, leading to suboptimal network performance.
Innovation Solution
A VNF placement system that uses a heuristic and mixed integer programming approaches to distribute service chain traffic across multiple nodes, balancing latency and resource utilization by routing traffic through different top-of-rack switches and instantiating VNFs on spare capacity, while considering heterogeneous server types and traffic rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If VNF placement systems assume fixed packet delays and impose latency constraints, then network performance can be maintained at acceptable levels, but resource utilization efficiency deteriorates leading to suboptimal network performance
Solution Approach 1:
The patent applies dynamics by transitioning from static fixed delay assumptions to dynamic latency modeling that adapts to actual network conditions. The system continuously monitors and adjusts latency parameters based on real-time traffic patterns and network state, enabling optimal resource utilization while maintaining service level agreements. This is evident in the iterative optimization process that refines VNF placement decisions based on observed performance metrics.
Solution Approach 2:
The patent utilizes parameter changes by modifying latency values from fixed constants to variable parameters that reflect actual network conditions. The system adjusts latency parameters dynamically based on traffic load, network topology changes, and service chain requirements, thereby improving both reliability and resource utilization efficiency simultaneously through data-driven parameter optimization.
2Quantity of substance
If service chain traffic is concentrated on fewer nodes, then the number of servers used is reduced, but congestion-induced latency increases deteriorating network performance
Solution Approach 1:
The patent applies local quality by distributing service chain traffic across multiple nodes with varying characteristics. Instead of uniform load distribution, the system places VNFs on specific nodes based on local conditions such as available capacity, network proximity, and resource availability. This creates optimized local placement decisions that reduce congestion on individual nodes while maintaining overall efficient server utilization.
Solution Approach 2:
The patent utilizes segmentation by dividing service chains into multiple VNF instances distributed across different nodes. Rather than consolidating all traffic on fewer servers, the system segments the service chain functionality across multiple locations, thereby reducing congestion-induced latency while achieving efficient use of the server pool through coordinated distribution.
3Loss of time
If VNFs are distributed across more nodes to reduce congestion, then latency is reduced, but the number of servers used increases reducing resource utilization efficiency
Solution Approach 1:
The patent applies preliminary action by pre-calculating optimal VNF placement configurations that balance latency reduction with resource utilization efficiency. The system performs advance optimization computations that identify placement strategies achieving low latency without excessive server usage, thereby avoiding the need to distribute VNFs across more nodes than necessary for performance requirements.
Solution Approach 2:
The patent utilizes feedback mechanisms to monitor actual latency and resource utilization metrics, then adjust VNF placement decisions accordingly. The system continuously evaluates whether the current distribution of VNFs across nodes achieves optimal balance between latency reduction and resource efficiency, making iterative adjustments to maintain this balance as network conditions change.
Data Source
AI summary
Example implementations relate to virtualized network function (VNF) placements. For example, VNF placements may include generating an initial mapping of a plurality of VNFs among a plurality of nodes of a network infrastructure, wherein the initial VNF mapping distributes each of a plurality of service chains associated with the plurality of VNFs to different top-of-rack switches. VNF placement may include generating an alternate VNF mapping of the plurality of VNFs among a portion of the plurality of nodes, wherein the alternate VNF mapping corresponds to a metric associated with node resource utilization and a particular amount of servers utilized by distributing the plurality of service chains according to the alternate VNF mapping. VNF placement may include placing the plurality of VNFs according to a selected placement from the initial VNF mapping and the alternate VNF mapping.