Zhu Hongwen

Traffic-based scheduling mechanism and performance analysis in IP/ATM router

We briefly present the design and software architecture of an IP/ATM multiple sub-network router. We establish the queue model of the IP process module and propose a traffic-based scheduling mechanism (AICS, adaptive incomplete cyclic service), on the basis of our prior work. We also provide a simulation study on the system throughput and packet delay jitter performance for both the AICS mechanism and IICS mechanism proposed in prior work. Simulation results show that, compared with IICS, AICS can provide much better packet delay jitter performance when high-speed sub-networks exist with bursty traffic and gives better fairness.

Harmonic retrieval in colored ARMA noise

We propose a new approach to harmonic retrieval in colored ARMA noise. A suitable filter is first used to remove all the sharp power spectrum peaks of the noisy observed process, then some kinds of cross correlation is employed to identify the noise characteristics. After filtering the noisy observed process with the identified noise characteristics again, SVD-TLS method can be applied to retrieve the harmonics. The proposed approach can be used to retrieve real-valued harmonic signals in colored ARMA noise with no restrictions on the phase coupling of harmonics and the distribution of the noise. Simulation examples show its effectiveness.

Region Tracking via HMMF in Joint Feature-Spatial Space

Region-based tracking in a temporal image sequence is described as a segmentation of current frame into a set of non-overlapping regions: the tracking regions and the non-tracking region. The segmentation is viewed to be a Markov labeling process. Based on the key idea of using a doubly stochastic prior model, the optimal estimation for the label field is found by the minimization of a differentiable function. We exploit the feature-spatial probabilistic representation of a region as the conditional distribution in the Bayesian framework, which makes our tracker robust to local deformation and partial occlusion. The continuity of the objective function leads to a much faster numerical implementation. Very promising experimental results on some real-world sequences are presented to illustrate the performance of the presented algorithm.

A routing algorithm in WDM mesh networks with grooming capacities

Most of the work in traffic grooming has been in the area of providing efficient network designs in Wavelength division multiplexing (WDM). The objective of these traffic rooming algorithms is to reduce the cost of overall networks. In this paper, a routing algorithm based on transiently chaotic neural network is proposed to solve the problem in WDM logical topology. The objectives of the routing algorithm are accomodating all traffic requirements and using less network resource. The simulation shows that the proposed algorithm is efficient in the routing selection, meanwhile the algorithm can use less network resource.

Transmission Considerations with QoS Support to Deliver Real-Time Distributed Speech Recognition Applications

Distributed speech recognition (DSR) applications have certain QoS (Quality of service) requirements in terms of latency, packet loss rate, etc. To deliver quality guaranteed DSR application over wirelined or wireless links, some QoS mechanisms should be provided. We put forward a RTP/RSVP transmission scheme with DSR-specific payload and QoS parameters by modifying the present WAP protocol stack. The simulation result shows that this scheme will provide adequate network bandwidth to keep the real-time transport of DSR data over either wirelined or wireless channels.

Neural network based scheduling for variable-length packets in Gigabit router with crossbar switch fabric and input queuing

A high-speed and effective packet scheduling method is crucial to the performance of Gigabit routers. The paper studies the variable-length packet scheduling problem in Gigabit router with crossbar switch fabric and input queuing, and a scheduling method based on neural network is proposed. For the proposed method, a scheduling system structure fit for the variable-length packet case is presented first, then some rules for scheduling are given. At last, an optimal scheduling method using Hopfield neural network is proposed based on the rules. Furthermore, the paper discusses that the proposed method can be realized by hardware circuit. The simulation result shows the effectiveness of the proposed method.

Analyzing and validating global fairness of TCP rate allocation

Network flow control is formulated as a global optimization problem of user profit in this paper. A general global optimization flow control model is used to study the global rate allocation characteristic of TCP congestion control mechanism. It is proved that with active queue management TCP rates tend to be allocated to maximize the aggregate of a user utility function U/sub s/ (called Unfairness) when the network reaches equilibrium. Compared with other fairness criteria, Us fairness is a more general conclusion because it considers the arbitrary topology and heterogeneous RTT case. A typical multi-bottleneck topology is used to validate Us fairness. Simulation results show the global rate allocation of TCP tends to accord with Unfairness in most scenarios.

Sender delay comparison of IP-based reliable synchronous collaboration communication

Synchronous collaboration is one of the most important fields of CSCW (computer supported cooperative work). IP multicasting is becoming the main trend of synchronous collaboration. But little of the current synchronous collaboration literature pays attention to the conditional superiority of IP multicasting over IP unicasting under a certain packet error rate. In this paper, we compare the sender delay performance of two reliable synchronous collaboration schemes which are based on IP unicasting and IP multicasting and show that IP multicasting is not always superior to IP unicasting in synchronous collaboration. In networks with low packet error rate and to a certain range of number of receivers, IP multicasting has the lower sender delay and superior scalability; while in networks with high packet loss rate and to a large user number, IP multicasting is inferior to the IP unicasting. In addition, we also conclude the communication principle which is to apply IP multicasting in low packet-error-rate environment and to apply IP unicasting in high packet-error-rate environment.

Narrowband interference rejection in stratospheric communication system using generalized TFD

Stratospheric communication system (SCS) is a new kind of wide area communication system that provides mobile wideband wireless access services. In order to be compatible with IMT-2000 and increase overall spectrum capacity, a CDMA overlay system is introduced into SCS. In this paper, the structure and characteristic of SCS are analyzed. Based on it, a detailed discussion about the narrowband interference rejection of SCS using a generalized time frequency distribution (TFD) is made. The simulation results of interference rejection are given at the end of this paper.

A new approach to VC routing in ATM networks

A new ant-algorithm-based routing approach is proposed for the VC routing problem with considering the comprehensive effect between the resource utilization and the load balance in ATM networks. In the approach, the backup paths are calculated first, and then an ant algorithm based on the ability of ants to find the shortest path between their nest and the food source during their searching food, is constructed to optimize the VC global route. Simulation results show that the proposed approach can realize VC routing effectively according to the current traffic states in the networks and the user-specified delay requirements.