Title :
Stochastic monotonicity and concavity properties of rate-based flow control mechanisms
Author :
Budka, Kenneth C.
Author_Institution :
AT&T Bell Labs., Holmdel, NJ, USA
fDate :
3/1/1994 12:00:00 AM
Abstract :
Using sample path comparisons, we study the stochastic monotonicity and concavity properties of the moving and jumping window flow control mechanisms, leaky bucket flow control mechanism, and the token bank rate control throttle to determine the effect control parameters have on throughputs and downstream congestion levels in packet switched networks. Results are developed without distributional assumptions on the packet arrival streams and make use of specially-tailored closed queueing networks. Performance comparisons between mechanisms are presented
Keywords :
directed graphs; packet switching; queueing theory; stochastic processes; closed queueing networks; concavity; jumping window flow control mechanisms; leaky bucket flow control mechanism; moving window flow control mechanisms; packet arrival streams; packet switched networks; rate-based flow control mechanisms; stochastic monotonicity; token bank rate control throttle; Adaptive control; Automatic control; Linear matrix inequalities; Mechanical factors; Optimal control; Programmable control; Riccati equations; Stochastic processes; Sufficient conditions; Testing;
Journal_Title :
Automatic Control, IEEE Transactions on