SDN Packet Scheduling with Priority Pushout for Edge Queues

The patent applies different scheduling rules to different packet priorities. High-priority packets receive preferential treatment (being pushed to the front of the queue) while low-priority packets follow standard FIFO ordering. This local differentiation in queue management resolves the contradiction by making the system easy to operate overall while reducing wait time for critical packets through priority-based repositioning.

The scheduling mechanism dynamically adjusts packet positions in the queue based on priority levels. When a high-priority packet arrives, it is moved to the front of the queue, changing the static FIFO order. This dynamic repositioning allows the system to adapt to varying traffic requirements without complex manual intervention, resolving the contradiction between operational simplicity and time efficiency.

The patent implements a pushout mechanism where low-priority packets at the tail of the queue are discarded when high-priority packets need to be inserted and the buffer is full. This selective discarding based on priority allows the system to maintain reliable delivery of critical packets without requiring excessively large buffers, thus avoiding the complexity of managing very large buffer structures while maintaining acceptable packet loss rates for important traffic.

The system changes the effective buffer allocation dynamically based on packet priority. Instead of a fixed uniform buffer, high-priority packets are granted preferential access to buffer space through the pushout mechanism. This parameter change in buffer allocation strategy allows the system to achieve high reliability for critical packets with moderate buffer sizes, avoiding the complexity of managing large uniform buffers.

The patent segments the network into multiple switches, each independently implementing the priority-based pushout scheduling algorithm. This segmentation allows the complex priority scheduling to be distributed across multiple simpler switch units rather than requiring a single complex centralized scheduler. Each switch handles its own queue management with relatively simple pushout logic, achieving high overall network performance while keeping individual switch algorithms manageable.

The pushout scheduling algorithm is designed as a universal mechanism that can be applied to multiple switches with the same basic logic. The same priority-based insertion and pushout rules work across different switches handling different traffic types (IoT data, video streaming, etc.). This universality allows the system to achieve high network-wide performance through a standardized, relatively simple algorithm that can be replicated across multiple switches without requiring complex switch-specific logic.