Credit token leaky bucket algorithm with fuzzy logic in ATM networks

This paper proposes a new model of leaky bucket (LB) called credit token leaky bucket (CTLB). Since the leaky rate of conventional LB is more than the negotiated mean cell rate, the performance of usage parameter control (UPC) depends on the burstiness of source and the statistical characteristics of cell stream passing through UPC. Therefore, lower performance, easily caused by a malicious user, may happen. In practical application fuzzy logic is chosen to be an artificial intelligent part which makes this mechanism work dynamically. Simulation results obviously show that the sensitivity of this proposed algorithm is better than that of conventional LB and is also close to that of ideal UPC. Furthermore, in the simulation results, CTLB returns fewer excessive cells going into ATM networks than the conventional LB does, while CTLB maintains the negotiated quality of service (QoS) for conformable sources, it passes cells without delay and still works with acceptable sensitivity.