Bounded Queue
A queue whose element count cannot exceed a fixed capacity, turning overload into an explicit wait, rejection, or load-shedding decision instead of unbounded memory growth.
A bounded queue imposes a finite upper limit on waiting jobs and thereby defines a capacity contract between producers and consumers. Queue capacity K, service capacity μ and arrival rate λ are considered together.
Load and latency
For a sustainable workload, long-run arrival rate must not exceed achievable service capacity. In a stable system, Little's Law gives L = λW: L is the average number of jobs within the system, λ the completed-job rate and W average time in the system. The relation does not by itself select a queue capacity; it connects work in progress with throughput and waiting time.
When a bounded queue reaches its limit, the admission policy may block producers, reject incoming work or apply priority-based load shedding. Per-job memory usage and deadlines are considered alongside element count. In real-time services, admission control must remain consistent with service rate and the time available to complete accepted requests.
Queue Stability in ASR Systems develops the model for speech-processing workloads.