| Popis: |
The extraordinary network challenges presented by high energy nuclear and particle physics experiments has created a need for network monitoring both to understand present performance and to allocate resources to optimize performance between laboratories, and the universities and institutes collaborating on present and future experiments. The resulting Internet end-to-end performance monitoring project is called PingER. The monitoring infrastructure reflects the wide geographical spread of the collaborations, and involves a large number of research and commercial networks. The architecture of the data acquisition and methodology of the analysis have evolved over several years, and are described here in their present state. The strengths and weaknesses of the project are reviewed, and the derived metrics are discussed in terms of their diagnostic functions. The observed short-term effects and long-term trends are reviewed, and plans for future developments are described. |
