Catálogo de publicaciones - libros

The multicast routing problem with quality of service (QoS) constraints is a key requirement of computer networks supporting multimedia applications. In order to resolve Qos multicast routing effectively and efficiently, an improved ant colony optimization (ACO) algorithm is proposed to resolve this problem.The core idea of improved ACO algorithm is mainly realized through pheromone local and global updating rule.Experimental results show that this algorithm can find optimal solution quickly and has a good scalability.

This paper proposes a new analytical model of PCS in torus in the presence of hot spot traffic pattern. Results from simulation experiments show close agreement with those predicted by the analytical model.

This work presents a new approach to software development framework design for parallel programming: the (). It adopts an incremental compilation process for OpenMP programming based on automatic detection of parallelism and user interaction for its calibration. GPE is extensible via plug-in modules providing new capabilities. It is an experimental OpenMP programming framework targeting shared-memory multiprocessors and clusters of PCs.

Current TCP congestion control can be inefficient and unstable in high-speed wide area networks due to its slow response with a large congestion window. Several congestion control proposals have already been suggested to solve these problems. In this paper, we propose a new variant of TCP for a high-speed network, which combines delay-based congestion control with loss-based congestion control. Our simulation results show that the proposed scheme performs better than the existing high-speed TCP protocols in terms of fairness, stability, and scalability, while providing TCP friendliness at the same time.

Streaming is a technology that enables either real-time or on-demand distribution of multimedia contents over network. Recently streaming technology has been applied onto applications, and many deployment tools for enterprise applications have been developed. Software streaming is a technology to provide software whichever users need on-demand in real-time by using streaming technology without downloading and installing a full package in advance before its use. Software streaming technology has many issues that are application load time, network fault-tolerant and etc. In this paper, we discuss issues for software streaming technology. Then, we propose a new SOD system based on functional streaming called system. Also, we present schemes to solve issues that are application load time and network fault-tolerant.

We investigate all possible combinations of re-ordering of read and write instructions and their effects on the correctness of programs that are designed for sequential consistency. With certain combinations of re-orderings, any program that accesses shared memory through only reads and writes and that is correct assuming sequential consistency, can be transformed to a new program that does not use any explicit synchronization, and that remains correct in spite of the instruction re-ordering. With other combinations of re-ordering, such transformations do not exist, without resorting to explicit synchronization.

Using the flow-conservation law, we could reduce the number of activated monitor agents used to monitor link bandwidth usage. In this paper, we address the problem of efficiently passive monitoring flow bandwidth based on flow-conservation, which could be reduced to weak vertex cover problem. And the weak vertex cover problem is NP-hard. We give an approximation algorithm with approximation ratio 2 to solve the problem. The effectiveness of our monitoring algorithm is validated by simulations evaluation over a wide range of network topologies.

Most heuristics for scheduling address asynchronous communication DAG, but they are not suitable for the synchronous ones. The proposed PRGSC algorithm avoids the deadlock that is caused by the synchronous communication, but also it can alleviate impact of synchronous communicating delay. Simulation shows that the PRGSC algorithm has better performance than the CASC algorithm which deals with the same type of problems.

Due to the heterogeneity of network processor architectures and constantly evolving network applications, it is currently a challenge to characterize the network processor workloads. In this paper, we formally model the task-level workloads of network processors as reactive dataflow process network (RDPN). RDPN is a suitable model of computation for formally describing the behaviors of packet-level parallel processing and event interaction with control point of network processors. We extend the expressive capability of RDPN by using three transformations (i.e., clustering, decomposing and duplicating) to analyze the model and support the further design space exploration of network processors.

This paper presents a new static data dependence analysis approach, Dependence Difference Inequality Test, which can deal with coupled subscripts for multi-dimensional array references for software pipelining techniques for nested loops. The Dependence Difference Inequality Test (DDIT) replaces direction vectors with dependence difference inequalities as constraints to variables in a linear system. The method presented in this paper extends the applicable range of the Generalized Lambda Test and seems to be a practical scheme to analyze data dependence. Experimental results show that the number of data independences checked by the DDIT algorithm is slightly smaller than that manually. It is also shown that our method is better than other traditional data dependence analysis methods without increasing time cost: it increases the success rate of the Generalized Lambda Test by approximately 14.19%.