Hybrid VOQ-OQ Queue Architecture for Switch Fabric Congestion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current queueing systems in network devices face challenges with scalability and head-of-line blocking, where Combined Input and Output Queuing (CIOQ) systems suffer from fabric congestion and limited throughput due to head-of-line blocking, while Virtual Output Queuing (VOQ) lacks output queue scaling as more packet forwarding engines are added.
Innovation Solution
Implementing a hybrid approach that combines elements of VOQ and CIOQ by using virtual output queuing at the egress side, allowing network packets to be enqueued in a virtual output queue and then looped back as metadata to an output queue, enabling increased queue scale and minimizing head-of-line blocking across the switch fabric.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If Combined Input and Output Queuing (CIOQ) is used to increase queue scale, then the number of supported packet forwarding engines increases, but head-of-line blocking occurs and throughput is limited
Solution Approach 1:
The patent segments the queueing system into two distinct parts: virtual output queues (VOQs) at the ingress side for initial packet buffering and classification, and output queues (OQs) at the egress side for final scheduling and transmission. This segmentation allows each queue type to specialize in its function, avoiding the head-of-line blocking problem that occurs in traditional CIOQ systems where all queuing functions are combined in a single structure.
Solution Approach 2:
The patent introduces a fabric intermediary that connects the VOQs to the OQs, enabling packets to be transferred from the ingress side to the egress side through a shared switching fabric. This intermediary structure allows independent management of input and output queues while maintaining high throughput, as packets can be scheduled and forwarded without being blocked by head-of-line conditions in a unified queue structure.
2Productivity
If Virtual Output Queuing (VOQ) is used to avoid head-of-line blocking, then throughput is maintained, but output queue scaling is limited as more packet forwarding engines are added
Solution Approach 1:
The patent adds a spatial dimension to the queueing architecture by distributing output queues across multiple egress packet forwarding engines. Instead of being constrained to a single VOQ structure, the system can scale the number of OQs by adding more egress engines, each with its own dedicated output queues. This dimensional expansion allows the system to maintain high throughput while simultaneously increasing the total queue scale to support more packet forwarding engines.
3Quantity of substance
If a hybrid approach combining VOQ and CIOQ elements is implemented, then queue scale and throughput are both improved, but system complexity increases
Solution Approach 1:
The patent implements a universal queue management framework where the same fabric switching mechanism handles traffic between VOQs and OQs, and the same scheduling algorithms can be applied at both ingress and egress sides. This multi-functionality allows the hybrid architecture to achieve both VOQ benefits (no head-of-line blocking) and CIOQ benefits (scalable queue capacity) without proportionally increasing complexity, as the same core components serve multiple functions in the system.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An apparatus for switching network traffic includes an ingress packet forwarding engine and an egress packet forwarding engine. The ingress packet forwarding engine is configured to determine, in response to receiving a network packet, an egress packet forwarding engine for outputting the network packet and enqueue the network packet in a virtual output queue. The egress packet forwarding engine is configured to output, in response to a first scheduling event and to the ingress packet forwarding engine, information indicating the network packet in the virtual output queue and that the network packet is to be enqueued at an output queue for an output port of the egress packet forwarding engine. The ingress packet forwarding engine is further configured to dequeue, in response to receiving the information, the network packet from the virtual output queue and enqueue the network packet to the output queue.