摘要： 网格监测是实现网格资源有效调度的基础，它采集、传输和分析各种资源的运行数据，由此实现性能评估、错误发现以及预测等，最终改善系统的性能。针对网格资源管理中的计算经济模型，提出了一个4 层的网格监测体系结构；设计了监测中的数据处理管理、监测数据交换；对体系结构中的模块组成和实现问题进行了讨论，在GRACE 上实现了原型系统。通过实验证明实现的监测体系结构是有效的。

关键词: 网格监测；计算经济；监测数据流；信息网关；体系结构

Abstract: Grid Monitor is the base of effective scheduling which collects, tramsmits and analyzes a variety of source data, so that it gives performance analysis, fault detection, prediction, and to improve system performance. Based on computational economy model, this paper presents a four-tier grid monitoring architecture, designs monitoring data exchanging model and data flow management, discusses modules structure and implementing problem in the architecture. Implementation of the new architecture prototype on GRACE proves that monitoring architecture is effective.

Key words: Grid monitor; Computational economy; Monitoring data flow; Data gateway; Architecture