NFV Flow Scheduling via Logical Field Calculations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Network Function Virtualization (NFV) solutions face challenges in efficiently distributing traffic to x86 processor cores, particularly in distinguishing and prioritizing user and application traffic, leading to suboptimal resource utilization and throughput due to software limitations and intensive TCAM-like lookups.
Innovation Solution
Implementing a method that classifies multiple flows from different customers and applications, distributing them to specific queues based on user and application priorities, using mathematical calculations to ensure flow order and efficient resource allocation across multiple processing resources, which can be achieved through hardware-based solutions like Field Programmable Gate Arrays (FPGAs).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all traffic from a singular port is directed to a singular core for processing, then flow order is maintained for all frames, but resource utilization is suboptimal and throughput is limited
Solution Approach 1:
The patent segments traffic flows into multiple classified flows based on user and application characteristics, then distributes these flows to different processor cores. This segmentation allows multiple cores to process different flows simultaneously, improving resource utilization while maintaining flow order within each classified flow through dedicated queue-core mappings.
2Productivity
If traffic is distributed based on Logical Interface classification (Port+VLAN, Port+IP destination), then resource distribution is improved, but multiple users in the same LIF still compete for the same CPU resources
Solution Approach 1:
The patent applies local quality by further classifying traffic within each Logical Interface based on application characteristics and user identifiers. This creates finer-grained flow classification that directs traffic from different users or applications to different processor cores, even when they share the same LIF. Each flow receives localized processing quality through dedicated core assignment, eliminating CPU resource contention while maintaining manageable classification complexity.
3Measurement precision
If further classification is performed in software-based NFV to distinguish multiple users or applications, then traffic distribution precision is improved, but TCAM-like lookups become intensive and software limitations are exposed
Solution Approach 1:
The patent replaces software-based classification mechanisms with hardware-based classification logic implemented in an FPGA. This substitution eliminates the performance limitations and high overhead of software-based TCAM-like lookups, enabling precise traffic classification and distribution while maintaining high-speed processing. The hardware implementation provides deterministic performance without the software overhead that plagues software-based NFV approaches.
4Productivity
If mathematical calculations are added to achieve controlled distribution of high-priority and low-priority traffic to different processors, then resource allocation optimization is improved, but system complexity increases
Solution Approach 1:
The patent performs preliminary classification of traffic flows based on user and application characteristics before distribution to processor cores. By pre-classifying flows and establishing deterministic distribution rules in hardware, the system optimizes resource allocation without requiring complex real-time mathematical calculations during packet processing. This preliminary action approach achieves efficient resource allocation while maintaining system simplicity through hardware-based rule enforcement.
Data Source
AI summary
Systems and methods of scheduling for Network Function Virtualization (NFV) on processing resources include receiving multiple flows from different customers and different applications; classifying the multiple flows to provide classified flows; distributing the classified flows to a plurality of queues; and providing each of the classified flows in the plurality of queues to the processing resources, wherein each individual classified flow is distributed to a same processing resource thereby maintaining flow order of the individual classified flow.