Most of the discussions about NSA data collection are devoid of technical facts. The media just likes to throw around the word “metadata” as if that means nothing to those of us who work all day with nothing other than metadata.
Here’s an article that doesn’t talk down to us, but explains how simple it is to replicate the HDFS nodes from Yahoo and Google data centers.
The problem seems to be that Yahoo and Google encrypt data in motion, but not data at rest. Would Accumulo solve the encryption problem for data at rest? However, Accumulo was originally developed for the NSA, who can likely break the encryption using the processing power of huge Hadoop clusters.
Source:
Cloud (IaaS) & Big Data 