Realization of wide-area distributed archival storage tolerable to large-scale failures

我们已经进行了基础研究，为实现广域分布式档案存储系统，将保持在互联网上的大规模故障下尽可能多的可访问性的基础。具体而言，我们集中在以下两个主题：（1）的方法，发现大规模的故障实际发生在互联网上的痕迹，通过调查每天快照的互联网拓扑结构，积累了大约8年，（2）的方法，预测数据传输吞吐量在云计算环境中的两个主机之间。

We have conducted fundamental research that are bases for realizing wide-area distributed archival storage system that will retain accessibility as much possible under large-scale failures in the Internet. Specifically, we concentrated on the following two topics :(1) methodology for finding the vestiges of the large-scale failures actually happened in the Internet by investigating daily snapshots of Internet topology that are accumulated for approximately 8 years,(2) methodology for predicting data transfer throughput between two hosts in cloud computing environment.