Catálogo de publicaciones - libros

Facing the limits of traditional tools of resource management within computational grids (related to scale, dynamicity, etc. of the platforms newly considered), new approaches, based on peer-to-peer technologies are emerging. The resource discovery and in particular the service discovery is concerned by this evolution. Among the solutions, a promising one is the indexing of resources using trie structures and more particularly prefix trees. The major advantages of trie-structured approaches is the capability to support search queries on ranges of values with a latency growing logarithmically in the number of nodes in the trie. Those techniques are easy to extend to multicriteria searches. One drawback of using tries is its inherent poor robustness in a dynamic environment, where nodes join and leave the network, leading to the split of the tree into a forest, which results in the impossibility to route requests. Within most recent approaches, the fault-tolerance is a prevention mechanism, often replication-based. The replication can be costly in term of resources required. In this paper, we propose a fault-tolerance protocol that reconnects subtrees a posteriori , after crashes, to have again a connected graph and then reorder the nodes to rebuild a consistent tree.

Random Early Detection (RED) is effective in decreasing losses on responsive flows but performs poorly with User Datagram Protocol (UDP) traffic since these do not react to congestion notification. However, it is important to address UDP flows’ requirements since UDP traffic forms a considerable part of Internet traffic. This paper shows how Value-Based Utility (VBU) with RED manages UDP flows. VBU allows packet dropping and queueing decisions to be more sensitive to UDP traffic packet loss requirements. Three RED variations incorporating VBU were developed. Results indicate adoption of VBU with RED enhances management of UDP traffic. This results in the equitable distribution of loss based on the UDP flows’ requirement.

In this paper, we envision a solution to the problem of capturing an intruder in a product network. This solution is derived based on the assumed existing algorithms for basic member graphs of a graph product. In this problem, a team of cleaner agents are responsible for capturing a hostile intruder in the network. While the agents can move in the network one hop at a time, the intruder is assumed to be arbitrarily fast in a way that it can traverse any number of nodes contiguously as far as no agents reside in those nodes. Here, we consider a version of the problem where each agent can replicate new agents. Hence, the algorithm start with a single agent and new agents are created on demand.

Recently, applications in which relational data is generated in a distributed and streaming manner have emerged from diverse domains. Processing queries on such data has become very important. In-network evaluation of a query is a technique in which the query is evaluated in the network without transferring all the data to a central location. So far, algorithms for in-network of evaluation of a single query have been proposed. They are not designed to exploit common computations across multiple queries. There is a need to develop techniques for efficient in-network evaluation of multiple queries. We consider the problem of in-network evaluation of multiple queries on relational data generated on a distributed network of machines. We present a novel algorithm based on an algorithm for dynamic regrouping of queries.

This paper presents an approach to optimal design of composite structures using Island Model Parallel Genetic Algorithm (IMPGA) with a probabilistic migration strategy and in conjunction with 3D Finite Element Method (FEM). The subject problem is computationally intensive and consumes large amount of computer space-time; an attempt has been made to spawn a variable number of processes at each node of IMPGA for FEM analysis. Comparison shows that integrated IMPGA-FEM module outperforms SGA-FEM module with respect to convergence as well as computational time significantly. The present study observes that the speed-up obtained from IMPGA-FEM module is better than the theoretical speed-up. It has also been observed that the incorporation of a probabilistic migration strategy in the IMPGA lead to a much faster and improved converged solution. Results show that for optimization with IMPGA there exists a minimum size of sub-population on each processor below which the performance deteriorates.

Emerging scientific and engineering simulations often require the coupling of multiple physics models and associated parallel codes that execute independently and in a distributed manner. Realizing these simulations in distributed environments presents several challenges. This paper describes experiences with wide-area coupling for a coupled fusion simulation using the Seine coupling framework. Seine provides a dynamic geometry-based virtual shared space abstraction and supports flexible, efficient and scalable coupling, data redistribution and data streaming. The design and implementation of the coupled fusion simulation using Seine, and an evaluation of its performance and overheads in a wide-area environment are presented.

The paper presents a tool that ports ScaLAPACK programs designed to run on massively parallel processors to Heterogeneous Networks of Computers. The tool converts ScaLAPACK programs to HeteroMPI programs. The resulting HeteroMPI programs do not aim to extract the maximum performance from a Heterogeneous Networks of Computers but provide an easy and simple way to execute the ScaLAPACK programs on such networks with good performance improvements. We demonstrate the efficiency of the resulting HeteroMPI programs by performing experiments with a matrix multiplication application on a local network of heterogeneous computers.

In this paper, a parallelizable computational technique for singularly perturbed reaction-diffusion problems is analyzed and implemented on parallel computer. In this technique, the domain is decomposed into non-overlapping subdomains, and boundary value problems are posed on each subdomain with suitable boundary conditions. Then, each problem is solved by the adaptive spline based difference scheme on each subinterval on parallel computer. Detailed theoretical analysis is provided to prove the convergence of the technique. To check the validity of the method, parallel implementation is performed on a numerical example and results are presented.

The software development process for many real-world applications requires many experts from different domains to cooperate. This is especially true for applications which are to be deployed in a distributed and heterogeneous environment such as the service-oriented Grid. In this paper, we present a collaborative synchronized Grid process creation environment for the service-oriented Grid, in which experts from different engineering and IT domains can interactively work on a single application process model, to quickly and efficiently design applications spanning multiple problem domains. An example application from the engineering domain of metal forming is presented and the enabling technologies are discussed.

A Bloom filter has been widely utilized to represent a set of items because it is a simple space-efficient randomized data structure. In this paper, we propose a new structure to support the representation of items with multiple attributes based on Bloom filters. The structure is composed of Parallel Bloom Filters (PBF) and a hash table to support the accurate and efficient representation and query of items. The PBF is a counter-based matrix and consists of multiple submatrixes. Each submatrix can store one attribute of an item. The hash table as an auxiliary structure captures a verification value of an item, which can reflect the inherent dependency of all attributes for the item. Because the correct query of an item with multiple attributes becomes complicated, we use a two-step verification process to ensure the presence of a particular item to reduce false positive probability.