Ke-Qiang Liao

Admission control and routing in ATM networks

Abstract Different approaches for effective bandwidth allocation in an ATM link are investigated. The results can be useful in constructing some simple admission models. A framework for flow control and routing in ATM networks is proposed. It is based on methods developed for circuit-switched networks.

A discrete-time single server queue with a two-level modulated input and its applications

A discrete-time, single-server system with a two-level modulated input is considered. The time axis is divided into equal-length slots, and the service time is deterministic and equal to one slot. The probability-generating function of the number of calls in the system is derived. A special case of this queuing system with a two-level Markov modulated Poisson input is studied in detail. Some numerical results concerning the mean and the variance of waiting time of an arbitrary call are given. The MMPP (Markov modulated Poisson process) approach is used to study some cases other than that of the packetized voice system. An alternative to the determination of the four parameters of the MMPP is also presented. This approach produces better results than the method presented for the cases with long burst length and high source peak rate (broadband integrated services digital network). This queuing model can be applied to performance analysis of a discrete-time ATM (asynchronous transfer mode) system and can be used as an approximation for a continuous-time MMPP/D/1 system.<<ETX>>

An approximate performance model for a multislot integrated services system

An approximate model (with finite or infinite waiting room) is presented for an integrated service system with three types of traffic: a first-offered narrowband traffic, an overflow narrowband traffic, and a wideband traffic. A narrowband call requires a single server, while the number of servers required to serve a wideband call is N. The blocked narrowband calls are lost while the blocked wideband calls are delayed in a finite or infinite waiting room. Based on two assumptions with regard to the characteristics of the system, the system is resolved by decomposition. The corresponding improvements in numerical efficiency as well as in computational storage requirements are significant enough to enable use of the model within network optimization algorithms. The model provides a very good approximation for the system performance, i.e. the blocking probabilities of the two narrowband traffics, the loss probability (in the case of finite waiting room), the probability of nonwaiting, and the average waiting time of wideband traffic. >

Flow control models for multi-service networks with delayed call set up

Admission control and routing in circuit-switched networks are investigated with two types of traffic: wideband with delayed call set up and narrowband operating in loss mode. Markov decision theory is used to construct a state-dependent routing policy. Optimality conditions for a load sharing routing are derived from the first-order Kuhn-Tucker equations. These models together with a direct routing scheme are used in a simulation study to determine the best approach for control of traffic with delayed call setup. State-dependent routing is found to be superior to direct routing and load sharing routing, although its operation is much more complex.<<ETX>>

Dynamic link bandwidth allocation in an integrated services network

The link bandwidth allocation problem in an integrated services network, where the types of calls are differentiated by the rates at which the requests for service are generated, the mean holding time, and the number of trunks requested, is considered. The decision to accept or reject a call for service may be based on the number of different calls in service to maximise (or minimize) a certain performance measure. This can be the throughput, the maximum blocking probability of any call type, or the revenue when different types have different call charging rates. An efficient and accurate approximation for an integrated services system where a dynamic trunk reservation policy is used is presented. The advantage of the dynamic trunk reservation over the steady trunk reservation is shown by numerical results.<<ETX>>

Effectiveness of leaky bucket policing mechanism

The authors consider the effectiveness of policing functions in controlling congestion in an asynchronous transfer mode (ATM) network. A method for measuring the effectiveness of a policing scheme is proposed, and analytical and simulation results are presented to show the effectiveness and limitations of a multiple leaky bucket policing scheme. The results indicate that statistical multiplexing with a policing scheme considerably improves the multiplexer efficiency compared to peak bandwidth assignment. An interesting result is that when a source is difficult to police, the gain from statistical multiplexing is large.<<ETX>>

Self-optimizing window flow control in high speed data networks

The problem of selecting the optimum window size in high-speed data networks is considered. A self-optimizing method is proposed to adapt the window size to network conditions. The scheme employs a cross-correlation technique for process identification with a random perturbation signal. Simulation results are presented for the isarithmic flow control to show the convergence rate and stability characteristics of the self-optimizing scheme.<<ETX>>

A heuristic approach for performance analysis of ATM systems

A heuristic approach to analyzing the performance of ATM (asynchronous transfer mode) systems is presented. It is based on a fluid approximation and a G/D/1 queue adjustment which removes the possible inaccuracy of the fluid approximation in light to moderate traffic. The performance analysis predicts the buffer content and packet delay distributions, which usually have stringent requirements. An infinite buffer is assumed in the analysis. For ATM networks, where the buffer overflow probability is small, the infinite-buffer model approximates the finite-buffer case well. A number of numerical examples show that this heuristic approach is very accurate over a wide range of system parameters.<<ETX>>