Link Aggregation Scheduler with Service Class Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current link aggregation methods, such as IEEE 802.3-2002, do not account for the services being transported over aggregated links, leading to unacceptable packet drops during bandwidth reductions without considering Class of Service (CoS), violating Service Level Agreements (SLA) when physical links fail.
Innovation Solution
A method where incoming packet flows are classified based on service classes and managed by a scheduler/shaper to ensure the global flow adheres to the aggregated link's bandwidth, using a distributor to signal bandwidth changes and prioritize packets based on CoS, queuing or dropping packets to maintain bandwidth within the logical link's capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If link aggregation is implemented without service awareness, then bandwidth is increased and operation is simplified, but packet drops occur without considering Class of Service during failures
Solution Approach 1:
The patent segments traffic into different service classes (e.g., gold, silver, bronze) with hierarchical priorities. Each service class is handled separately through dedicated queues and scheduling policies, allowing the system to preserve high-priority traffic while dropping lower-priority traffic during bandwidth constraints, thus maintaining SLA compliance.
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously monitors link status, bandwidth availability, and service class requirements. Based on this feedback, the scheduler dynamically adjusts packet selection and queue management to ensure that service-aware drop policies are applied, preventing SLA violations during failures.
2Device complexity
If standard link aggregation is used, then device complexity is reduced, but the system cannot differentiate or prioritize traffic during bandwidth reductions
Solution Approach 1:
The patent introduces dynamic service class differentiation within the aggregated link framework. The system can dynamically adjust scheduling parameters, queue priorities, and drop policies based on current bandwidth availability and service requirements, allowing adaptability without increasing fundamental aggregation complexity.
Solution Approach 2:
The patent makes the link aggregation system multi-functional by enabling it to handle both standard traffic aggregation and service-aware traffic management simultaneously. The same aggregation infrastructure supports multiple service classes with different priorities, making the system versatile while maintaining operational simplicity.
3Quantity of substance
If bandwidth is reduced due to physical link failures, then available capacity decreases, but packet drops occur without regard to service class priorities
Solution Approach 1:
The patent applies local quality by assigning different treatment qualities to different service classes. High-priority service class packets receive preferential treatment with lower drop probabilities, while low-priority packets are more readily dropped during bandwidth reductions. This localized quality differentiation ensures that critical services maintain performance even when overall bandwidth is reduced.
4Reliability
If service-aware scheduling is implemented, then Quality of Service is improved, but system complexity increases
Solution Approach 1:
The patent implements partial service-aware scheduling by focusing on the most critical aspects of QoS management. Rather than implementing complete end-to-end service management, the scheduler/shaper applies service-class-based prioritization and drop policies at the aggregation point, achieving significant QoS improvement with moderate complexity increase.
Data Source
Figure 1~2
Figure 3
AI summary
A method and a transport scheme for packets of traffic over a logical link made up of the aggregation of several physical links (16) connecting a transmitting side to a receiving side in which flows of incoming packets are sent to a scheduler/shaper (12) which selects therefrom packets for creating a global flow of packets falling within the bandwidth offered by the logical link on the basis of the bandwidth capability offered by the logical link. A distributor (13) distributes the global flow over the plurality of physical links (16) making up the logical link, oversees the physical links and sends to the scheduler/shaper signalling of a bandwidth decrease caused by failures of one or more physical links. The scheduler/shaper (12) is arranged to allow automatically for the logical link bandwidth variations while selecting the packets from the various incoming queues depending on the associated service class so as to cause the global flow to fall within the aggregated logical link bandwidth. This involves more queuing of packets and their possible drop according to the queue management criteria implemented, in the queues associated with the lowest service class.