An interesting spin on an old technique brought to Hadoop

http://eprint.iacr.org/2012/398.pdf

"...Retrieval of previously outsourced data in a privacy-preserving manner is an important requirement in the face of an untrusted cloud provider. PIRMAP is the rst practical PIR mechanism suited to real-world cloud computing.  In the case, where a cloud user wishes to privately retrieve large les from the untrusted cloud, PIRMAP is communication ecient. Designed for prominent MapReduce clouds, it leverages their parallelism and aggregation phases for maximum performance. Our analysis shows that PIRMAP is an order of magnitude more ecient than trivial PIR and introduces acceptable overhead over non-privacy-preserving data retrieval. Additionally, we have shown that our scheme can scale to cloud stores of up to 1 TB on Amazon's Elastic MapReduce service..."

​