Coreografia de fluxos de trabalho científicos para computação em grades

Abstract

Computational Grids are a broadly used distributed computing environment. They are used mainly in academic environments and scientific research and have improved the collaboration of research groups in projects deployed in national, continental and global levels. Among existing Grid environments, those seeking to create or improve an environment for research in medical sciences are being given a lot of attention. An important example of such a type of Gridenvironment is caGrid, which was created to improve scientific research on topics related to cancer. In a Grid environment like caGrid, communities arebuilt and new needs arise. One of the main needs of caGrid is the development of a tool for helping researchers to use scientific workflows, which are complex sequences of tasks built on top of data and services available in theGrid. To meet the need of the community of users of caGrid for a tool to helpin the use of scientific workflows, the author of this dissertation developedthe caGrid choreography System (caOS). The developed system enables a scientistto create, execute and monitor a scientific workflow based on a description ofthe workflow in a well-defined format. The main results of this work are: theuse of choreography for executing scientific workflows, the development of security mechanisms integrated to those already existing in caGrid and aneasily extensible architecture of caOS.