Arbitrated Bid Granting in Switch Crossbars

The patent segments the switch architecture into multiple independent crossbars instead of using a single bus or crossbar. Each crossbar operates independently to handle specific traffic flows, allowing the system to scale bandwidth by adding more crossbars without proportionally increasing overall complexity. This segmentation enables Terabit-per-second throughput by distributing traffic across multiple parallel paths.

The patent introduces a new dimension to the switching architecture by adding multiple crossbars operating in parallel, transforming the single-plane switching fabric into a multi-plane architecture. This dimensional expansion allows simultaneous handling of multiple traffic streams with different QoS requirements, achieving Terabit bandwidth through spatial parallelism rather than increasing the complexity of a single crossbar.

The patent implements preliminary classification of information packets into channels based on QoS parameters before the arbitration process. This pre-processing organizes traffic into priority-based channels, so that during arbitration, high-priority packets are automatically granted access first. This preliminary action eliminates the need for complex real-time priority calculations during arbitration, reducing processing time while ensuring fair QoS-based access distribution.

The arbitration mechanism allows each channel to self-manage its packet transmission based on pre-assigned priorities and available bandwidth. Channels with higher QoS requirements automatically gain preferential access without requiring continuous centralized arbitration intervention. This self-service approach reduces arbitration overhead and processing time while maintaining fairness through QoS-based priority management.