Processing Information Without Learning It
Wednesday, February 09, 2011 2:30 PM - 3:30 PM EST
Although utility-computing technologies like cloud computing can yield substantial economic, social, and scientific benefits, there are impediments to achieving their full potential. One of the impediments is a reluctance to disclose information, for fear of losing control over its subsequent dissemination and usage. Moreover, laws often forbid the disclosure of certain kinds of information (e.g., health, financial), or strictly regulate the form and timing of that disclosure. We review security technologies that can mitigate this problem, and that allow data owners to enforce their approved purposes on their data. We consider both the computational and the storage aspects of this problem. This includes computational outsourcing, in which weak clients use powerful remote servers to carry out intensive computational tasks without revealing to the servers anything about the data or the computed answers; and storage outsourcing, in which storare-limited clients use remote servers to store/search/manipulate massive data without revealing to the servers anything about the data or the queries and updates on it.