Service Scheduling Device Token Threshold Configuration

The patent segments the single shared token bucket into multiple priority-specific token buckets, where each priority level has its own token bucket with dedicated token thresholds. This segmentation allows high priority queues to have guaranteed token availability independent of low priority queue transmissions, resolving the contradiction by reducing device complexity through structured organization while eliminating delay and jitter for high priority services.

The patent applies local quality by assigning different token threshold characteristics to different priority levels. High priority queues are allocated token buckets with lower thresholds that are easier to replenish, while low priority queues have higher thresholds. This localized differentiation ensures that each priority level receives appropriate resource allocation, preventing low priority long packets from causing high priority delay and jitter while maintaining overall system efficiency.

The patent implements dynamic token threshold management where each priority level has its own token bucket with dynamically adjustable thresholds. The system can dynamically allocate tokens based on current network conditions and priority requirements, allowing low priority queues to transmit when tokens are abundant while ensuring high priority queues always have access to sufficient tokens. This dynamic approach resolves the contradiction by enabling flexible resource allocation that optimizes both productivity and timing performance.

The patent changes the parameter of token thresholds from a single unified value to multiple priority-specific values. By setting different token thresholds for different priority levels (with high priority having lower thresholds), the system allows low priority services to achieve good productivity when tokens are available while ensuring high priority services experience minimal delay and jitter through their dedicated lower thresholds that are replenished independently.

The patent segments the token restoration process into priority-specific streams by providing separate token buckets for each priority level. When tokens are consumed by long packets in low priority queues, only the low priority token bucket is affected, while high priority token buckets maintain their token levels independently. This segmentation eliminates the waiting time for high priority services by decoupling their token restoration from low priority queue activity.

The patent applies preliminary action by pre-allocating dedicated token buckets to each priority level with predetermined threshold values. High priority queues are pre-configured with token buckets that have lower thresholds and are replenished with higher priority, ensuring that tokens are available in advance before low priority queues need to transmit. This preliminary preparation prevents high priority services from having to wait for token restoration after low priority packet transmission.