Catálogo de publicaciones - libros

A Grid is a family of technologies for dynamically and opportunistically provisioning computing power from a pool of resources. Some experts believe that Grids are the next stage in the evolution of distributed systems. Enterprise-level Grids are beginning to be deployed, and researchers are investigating a number of novel Grid applications and services in Grid testbeds. Production-grade commercial Grids are expected to be large and widely-distributed systems. Grids are enormously complex computing systems. They have large numbers of geographically dispersed resources at their disposal, and they lack some characteristics of earlier computing systems that simplified the analysis of those systems — e.g., homogeneous and closely-coupled components, low-latency communications paths, global state maintenance, and deterministic controls. Performance analysts have very little understanding of the dynamic behavior of Grid components at different layers and at different time scales, and of the complex interactions among Grid components — functionally, spatially, and temporally. We describe an architecture for a scalable, modular, configurable, plug-and-play, “Grid-in-a-lab” tool for modeling Grids, and for analyzing the performance of applications and middleware services offered on Grids. The tool addresses all of the components that are relevant to Grid performance — hardware, middleware, infrastructure, services, applications, traffic — and the complex interrelationships among the components. It uses a novel combination of modeling techniques — e.g., analytical models with closed form expressions, discrete event simulation, emulators for protocol stacks and resource scheduling, rare event simulation methods, and measurement-based optimization. The tool’s components are modular and configurable so as to provide different degrees of fidelity over a range of time scales, and to provide insight about Grid dynamics and interactions among various components, applications, and services.

The basic concept and evolution of the network simulator (NS) is discussed MPLS Network Simulator (MNS) and Wireless Mobile Networking are described with emphasis on conceptual model and architecture. Current NS compatibility issues and shortcomings are considered and a potential development plan is proposed. NS enhancement strategies in terms of enriching various modules with MPLS and HMIPv6 functionalities are proposed and novel schematic embedding algorithms are shown.

Integrating simulation of communication networks with a logistical networks simulation is investigated taking advantage of object-oriented languages’ features and design patterns. In this paper we present asimulator for logistic processes and the way how it is extended to enable combined simulation of logistical and communication networks. The aim of this integration is to measure the influence that an improved communication has on logistical transport processes.

As deployment of smart vehicles becomes more widespread, the application of system evaluation methodologies to networked in-vehicle computing systems becomes increasingly more important. Such methods are essential to understand system behaviors as well as to assess alternate approaches toward realizing a rich variety of computing and information services to travelers. We describe our experiences in evaluating vehicular networks using analytic analysis, simulation, and field experiments. Each evaluation method presents its own challenges and offers different strengths and weaknesses. The context for these investigations concerns the exploration of the use of vehicle-to-vehicle and vehicle-to-infrastructure communication to disseminate information in vehicular networks.

Burst switching appears to be the most promising technique for meeting transperancy and bandwidth-on-demand requirements in wavelength division multiplexed, ultra high capacity, backbone optical networks. Modeling and simulation of such networks present challenges which are rather difficult, and in extreme cases impossible to meet using analytical or discrete event approaches. This is due to the absence of sufficiently accurate models for the former, and due to the required exessively large computational resources for the latter approach. In this paper a different modeling approach is presented, which under the assumptions of stationarity of distributions allows for implementation of Monte Carlo sampling for calculating overall burst loss probability performance of JIT protocol in optical burst switched networks. Simulation results are presented for NSFNet and Pan-European Network, which demonstrate the applicability of the proposed approach.

A Markov chain based modeling and simulation approach is presented for estimating performances of radio channels and modulation schemes. The model parameters, i.e. the number of Markov states and transition matrix elements can be derived in a statistically consistent way, from measured data. Once the model parameters are set time series generated via simulation can be used for performance prediction. Example outputs are given which demonstrate that the predicted results for BER in a particular radio channel are in good agreement with the measured average value.

The article describes the philosophy of component approach to system design, illustrated on the optical transmission network example. The approach is based on the object oriented paradigms of Cosmos and Nyx tools. Cosmos serves as the support for system and component development, behavior description, and simulation and analytic calculation development. Nyx in turn enables description of optimization procedures using general heuristic search techniques. The combination of these two development environments and their component approach enables reusability and shortening of the new application development time.

This paper addresses the basic concepts of the packet switching in the optical domain and describes an analytical approach to evaluate the end-to-end performance of networks employing slotted (fixed length) optical packets. Thus, for a given topology and traffic matrix, the end-to-end cell loss ratio is analytically computed assuming an uncorrelated traffic. A network dimensioning method relying on this approach is also presented.

Modeling and simulation (M&S) of network-centric military systems poses significant technical challenges. A network-centric military system (also known as network-centric operations, network-centric warfare or FORCEnet) is a system of systems aligning and integrating other systems such as battlefields, computers, databases, mobile devices, people (users), processes, satellites, sensors, warriors, shooters, and weapons into a globally networked distributed complex system. Characteristics of a network-centric military system are described using a layered architecture. Challenges for M&S of network-centric military systems are presented. The paper focuses on the quality assessment challenge and advocates the use of a quality model with four perspectives: product, process, project, and people. A hierarchy of quality indicators is presented for network-centric military system M&S. An approach is described for conducting collaborative assessment of M&S quality using the quality indicators.

The process of accrediting a model can roughly be divided into first gathering and evaluating results of conducted verification and validation (V& V) activities, and then aggregating these (raw) results into a total value: the accreditation decision. Especially the second part is not trivial, because results of V& V activities are quite manifold and at first sight not comparable with each other. Consider for example the evaluation of subject matter expert statements in contrast to real numbers as an outcome of statistical tests. Classical decision analysis is one possibility to support the accreditation decision by using a structured, scientifically justified approach. For that purpose, mainly the concept of value tree analysis is used. In previous papers, the current state-of-the-art concerning decision analytic methodologies supporting VV&A was reviewed. Also a critical examination, describing gaps and deficiencies was conducted. A concept was drafted, how fuzzy value tree analysis, a branch of decision analysis incorporating fuzzy set theory, can be used to overcome some of these deficiencies, as there are: First, modeling subject matter experts statements, second, quantifying not only the value, but also the knowledge resp. ignorance about the model under study, and finally, being able to distinguish between compensational and non-compensational attributes in the value tree. Fuzzy value tree analysis consists of the steps establishing the value tree, assigning values to attributes, aggregating the attribute values, and interpreting the result. A raw draft of the concept, as well as a definition of attribute values in a fuzzy value tree analysis was dealt with in previous publications. The paper at hand takes these publications as a starting point and presents a concept how attribute values in a fuzzy value tree analysis can be aggregated, i. e. the aggregator is defined. Starting from establishing requirements and constraints, these postulations are transformed into mathematical properties the aggregator must adhere to. After scanning the literature for possible fuzzy set theoretic operators, it is proven that the defined aggregator fulfills the mathematical properties established and therefore all requirements and constraints. The paper closes with illustrative examples.