Port-Based Fairness Protocol for Daisy Chain Network Bandwidth Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Daisy chain network configurations often lead to unfair bandwidth distribution, particularly affecting the first network element, known as the 'parking lot problem', which existing solutions like per-flow queuing and per-flow accounting/dropping may not adequately address due to high costs and increased latency.
Innovation Solution
A fairness protocol that monitors and compares bandwidth from multiple ingress ports, giving precedence to traffic from ports with lower bandwidth, ensuring fair bandwidth allocation regardless of the number of ports, using methods like maintaining bandwidth tables and statistical analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If per-flow queuing is used to address the parking lot problem, then bandwidth fairness is improved, but device complexity and processing cost increase significantly
Solution Approach 1:
The patent segments the daisy chain network into multiple groups, with each group having a representative port. Instead of managing per-flow queuing across all ports, the system creates separate virtual output queues (VOQs) for each group representative, significantly reducing the number of queues and scheduling complexity while maintaining bandwidth fairness within each group.
Solution Approach 2:
The patent introduces group representative ports as intermediaries between individual ports and the core switching fabric. These representatives aggregate traffic from multiple ports and participate in arbitration, simplifying the scheduling mechanism while ensuring fair bandwidth allocation. The representatives act as mediators that reduce the complexity of direct port-to-fabric scheduling.
2Reliability
If per-flow accounting and dropping is used to address the parking lot problem, then bandwidth fairness is improved, but latency increases due to queuing at each node
Solution Approach 1:
The patent performs preliminary bandwidth arbitration at the ingress side before packets enter the switching fabric. By determining the winning group in advance and pre-establishing the transmission schedule, packets avoid repeated queuing and arbitration delays at each intermediate node, significantly reducing end-to-end latency while maintaining fairness.
Solution Approach 2:
The patent extracts the arbitration and scheduling function from individual network elements and consolidates it at the ingress side. This removes the need for packets to be queued and re-arbitrated at each daisy-chained node, eliminating the cumulative latency that would otherwise occur as packets traverse multiple elements.
3Reliability
If per-flow accounting and dropping is used to address the parking lot problem, then bandwidth fairness is improved, but maintenance overhead increases due to per-flow drop counters
Solution Approach 1:
The patent merges the accounting function from the individual port level to the group representative level. Instead of maintaining per-flow drop counters at each port, the system maintains accounting information at the group representative, reducing the number of counters and simplifying maintenance while still enabling fair bandwidth allocation through group-based arbitration.
4Productivity
If traditional arbitration is used in daisy chain configuration, then the first network element gets most bandwidth, but fairness among multiple ports is poor
Solution Approach 1:
The patent introduces asymmetry in the arbitration mechanism by giving different priorities to different groups based on their current bandwidth usage. The arbitration is not symmetric across all ports but instead adapts to each group's needs, with groups having lower bandwidth allocation receiving higher priority in subsequent arbitration rounds, thereby achieving fairness despite the asymmetric treatment.
Solution Approach 2:
The patent implements dynamic bandwidth allocation where the arbitration priorities and group compositions can change based on current network conditions and traffic patterns. This dynamic approach allows the system to adapt to varying load conditions and maintain fairness over time, rather than using a static arbitration scheme that would consistently favor certain ports.
Data Source
AI summary
Methods and apparatuses for providing a fairness protocol in a network element are disclosed herein. In accordance with the disclosed fairness protocol, the average bandwidth of traffic sourced from each of a plurality of ingress ports is monitored. The largest bandwidth of traffic sourced from a port within a first group of ingress ports (e.g., ingress ports of a network element) is identified and compared to the largest bandwidth of traffic sourced from a port within a second group of ingress ports (e.g., ingress ports of one or more network elements communicatively connected to the network element). In order to fairly allocate bandwidth when transmitting traffic that is sourced from the first and second groups, precedence is given to traffic flowing from the group associated with the identified port sourcing the lower bandwidth of traffic.