Network Switch Queueing for Asymmetric Optical Bandwidth Allocation

The patent applies local quality by assigning different transceiver capacities to different network nodes based on their specific roles and requirements. Primary nodes (network switches) are equipped with high-capacity transceivers to handle aggregate traffic from multiple secondary nodes, while secondary nodes (server computers) use lower-capacity transceivers matched to their individual data rates. This differentiated approach optimizes overall network throughput without requiring all nodes to use expensive high-capacity transceivers, thereby reducing device cost and network complexity.

The patent segments bandwidth allocation into two distinct directions: downstream (primary to secondary nodes) and upstream (secondary to primary nodes). For downstream traffic, pooled bandwidth allocation is implemented where the primary node aggregates data from multiple secondary nodes and transmits over a high-capacity transceiver, improving bandwidth utilization efficiency. For upstream traffic, dedicated bandwidth allocation is used where each secondary node has its own allocated capacity, ensuring reliable one-to-one correspondence between transceivers and data rates. This segmentation resolves the contradiction by applying appropriate allocation strategies to different traffic directions.

The patent applies asymmetry by treating downstream and upstream bandwidth allocation differently. Downstream traffic (from primary to secondary nodes) uses pooled bandwidth allocation where multiple data streams are aggregated and transmitted over a single high-capacity transceiver, achieving high bandwidth utilization efficiency. Upstream traffic (from secondary to primary nodes) uses dedicated bandwidth allocation with one-to-one transceiver correspondence, ensuring reliable data transmission. This asymmetric approach allows the network to optimize for both efficiency and reliability in different directions, avoiding the need for dedicated allocation everywhere which would reduce overall efficiency.

The patent applies parameter changes by varying transceiver capacity parameters according to the specific requirements of different network nodes and traffic directions. Primary nodes use high-capacity transceivers (first maximum throughput) to handle aggregate downstream traffic, while secondary nodes use lower-capacity transceivers (second maximum throughput) matched to their individual data rates. This parameter differentiation ensures that data transmission speed is optimized where needed (downstream from primary nodes) while avoiding the excessive cost and energy consumption of deploying high-capacity transceivers at all nodes. The system dynamically adapts transceiver parameters to match actual traffic demands.