Ethernet Bonding Processor Frame Reordering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies for bonding multiple physical Ethernet interfaces into a single logical interface face challenges in preserving the order of Ethernet frames due to varying transmission latencies, which can lead to mis-ordering, especially in point-to-multipoint networks where legacy endpoints may not support bonding, and existing solutions increase complexity in Medium Access Control logic.
Innovation Solution
A bonding processor that inserts sequence numbers into Ethernet frames, distributes them based on load distribution criteria considering frame lengths, and reserves synchronized timeslots to maintain frame order and balance load across physical interfaces, ensuring backwards compatibility with existing protocols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If Ethernet frames are transmitted across multiple physical interfaces without sequence numbering, then transmission speed is improved, but frame order preservation deteriorates
Solution Approach 1:
Sequence numbers are inserted into Ethernet frames before distribution across multiple physical interfaces. This preliminary action enables the receiving bonding processor to correctly reassemble frames in their original order, solving the frame ordering problem while maintaining high transmission speeds across bonded interfaces.
2Reliability
If sequence numbering is implemented to preserve frame order, then frame order preservation is improved, but device complexity increases
Solution Approach 1:
The bonding processor functionality is segmented into distinct modules: sequence number insertion, frame distribution across physical interfaces, and reassembly at the receiver. This segmentation isolates the complexity to specific components while keeping the overall system manageable and the Medium Access Control layer simple.
3Device complexity
If load distribution is based on simple criteria, then device complexity is reduced, but productivity decreases
Solution Approach 1:
The load distribution mechanism uses frame length as a dynamic parameter to determine which physical interface receives each Ethernet frame. This simple parameter-based approach balances the load across bonded interfaces effectively, improving productivity without introducing complex algorithms or increasing device complexity.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A bonding processor (133) for backwards compatible bonding of plural physical interfaces (121, 122, 123, 124) carrying Ethernet frames or packets into a single logical interface carrying Ethernet frames or packets allows bonded transmission in a network wherein the physical interfaces (121, 122, 123, 124) carrying Ethernet frames or packets are also used for non-bonded transmission. The bonding processor (133) comprises: - means (191) for inserting a sequence number (206; 307) in Ethernet frames or packets to thereby generate sequence-numbered Ethernet frames or packets (200; 300); - means (192, 193) for distributing the sequence-numbered Ethernet frames or packets (200; 300) across plural physical interfaces (121, 122, 123, 124) carrying Ethernet frames or packets according to a load distribution criterion that considers the respective lengths of the sequence-numbered Ethernet frames or packets (200; 300); and - means (195) for reserving substantially simultaneous timeslots (431, 451, 471, 491) on the plural physical interfaces (121, 122, 123, 124) carrying Ethernet frames or packets for bonded transmission of the sequence-numbered Ethernet frames or packets (200; 300).