An agent architecture for network support of distributed simulation

Abstract not found

In Web Services designs classical optimization techniques are not applicable. A possible solution to guarantee critical requirements is the use of an autonomic architecture, able to autoconfigure and to auto-tune. This study presents MAWeS (MetaPL/HeSSE Autonomic Web Services), a framework whose aim is to support the development of self-optimizing predictive autonomic systems for Web service architectures. It adopts a simulation-based methodology, which allows to predict system performance in different status and load conditions. The predicted results are used for a feedforward control of the system, which self-tunes before the new conditions and the subsequent performance losses are actually observed.

This paper is available online at www.jtaer.com DOI: 10.4067/S0718-18762010000200005

Abstract not found

Modern communication networks are reaching a high level of complexity. As their global dimensioning and performance evaluation are very hard tasks, a multi-layer approach is generally applied to divide them in smaller problems easier to tackle. Within this approach, tools only address one or two network layers. To perform an enhanced analysis encompassing more layers, several tools, not necessarily available in the same place when the problem arises, are often required. This paper shows how to allow a remote access to various existing tools using Web Services paradigm. In this way, software components modelling network elements, layers, or computing specific functions are turned into Web Services, which are available to remote users over the internet. Accessed through a simplified interface, they can be used sequentially or in parallel to solve a specific task. This dispenses the user to perform a local implementation, allows better code reuse, and offers an easy way to confront results from distinct models. A tool accessed via a web service can be beneficial for either research or educational purposes. Moreover, making tools available on the web increases credibility and visibility of their authors. Calls to Web Service require transmitting and retrieving all

as simulation input, and they strongly influence the results. Thus, their usage and the usage of their generator need to be taken care of very well. Qualified generators are available on the web as source code or libraries. However, they require an additional middleware to adapt them to the running environments, and this can lead to misuses. This paper proposes Pseudo Random Number Generators embedded inside a Web Service for the use in simulations. This service, while offering a unique interface accessible from any platform, eases the important task of retrieving correct pseudo random numbers. The general architecture of the service is presented, as well as a reference implementation. The performance of the Web Service generator is compared to the performance of local generators.

Abstract — Data management for linked (or coupled) simulation models can be a challenging task when deploying to grid environments. In cases where the linked models conform to a standard interface for data input and output, generalpurpose data providers can be used to supply data to the models from online sources, reducing the complexity of the deployment. We have developed a data provider component that conforms to the Open Modeling Interface (OpenMI) that is suitable for use on computational grids. Through the application of three techniques, caching, prefetching, and pipelining, the component efficiently retrieves data from standards-based web services and delivers the data to OpenMI-compliant models. Each technique resulted in varying performance improvements both within a single simulation and across multiple simulations concurrently executing on a cluster. In this paper we report on the design of the component and the evaluation of its performance.

Abstract. Successfully communicating an organisation’s goals and rules in such a manner that their impact on IT support can be analysed and discussed, presents a major challenge. The most significant problem is relating the changes desired in a semantically consistent and understandable manner and then reflecting the potential impact of those changes on the organisational structure and the business processes carried out within that organisation. This research investigates a simulation-based communication tool that employs a semantically driven natural-language component to capture an organisation’s needs. These needs are translated into a simulation, through the use of semantic web services, of a business process that represent the evolution of the organization’s IT infrastructure and policies. Through an iterative communication loop between the business executive, IT architect and staff, the aim of the simulation is to accurately represent and specify the IT changes necessary to support the organisation’s needs. An initial scenario-based evaluation of this idea that we have undertaken indicates a desire among executives for such a simulation tool. This paper introduces the problem, architecture and case study, followed by a discussion on an evaluation that is in progress at the National College of Ireland with results expected by the end of March.