Stochastic modelling of maintaining specified QoS constraints in discrete-time domain

Quality of service (QoS) is of extreme importance to accommodate the increasingly diverse ranges of services and types of traffic in the Internet, and maintaining the delay at a specified value is a key QoS requirement. In this paper, a two-state Markov-modulated Bernoulli process (MMBP) is set up as our source model. To satisfy some specified delay, we present an analytical model for a discrete-time finite capacity queue, which incorporates two thresholds (L₁ in state-one and L₂ in state-two) in the system so that the source can be blocked. If the number of packets in the system reaches the threshold, then the source stops sending the packets, otherwise the source operates normally. That can be considered as an implicit feedback from the queue to the arrival process. In addition, when the arrival process is in a specific state, this information is used to adjust the thresholds L₁ and L₂ so that the mean delay through the queue is maintained at some specified value to satisfy QoS requirements.