Pervasive Grids with Autonomic Capabilities

一个网格计算系统，汇集了大量的异构资源能够应该能够连续运行，没有太多的干预，由人类操作员。这项工作的目的是开发技术和工具，用于监测和预测网格核心结构的行为。在这种情况下，？核心？结构是一个大型或广域网，或这些网络的集合。监控过程将向网格结构中的其他层提供重要信息（流量、当前和可能的未来故障、故障、拓扑变化等），以实现网格的高效和一致的操作。这是网格计算中的一个重要研究问题 因为在传统的高性能计算环境中或多或少有效的传统假设在网格上被打破了。

A grid computing system that brings together a multitude of heterogonous resources able should be able to function continuously without much intervention by a human operator. This work aims at developing techniques and tools for the monitoring and prediction of the behaviour of the core structure of a grid. In this case, the ?core? structure is a large- or a wide-area network, or a collection of such networks. The monitoring process will feed into the other layers in the grid fabric important information (traffic, current and possible future congestions, failures, topological variations, etc) to enable the efficient and consistent operation of the grid. This is an important research problem in grid computing since traditional assumptions that are more or less valid in conventional high-performance computing settings break down on the Grid.