Efficient proactive secret sharing

The secure storage of long-lived sensitive data is constantly growing in its relevance due to the ever increasing digitization of documents. One very important challenge of this research field is to provide confidentiality for the stored data even in the long term. The only known approach to achieve this, as required, for instance, for medical records, is to use proactive secret sharing. However, all currently known schemes suffer from being inefficient. They require information-theoretic secure communication channels between any two shareholders and between the client and each shareholder and come with a high communication complexity. Thus, this work addresses the scenario where only a subset of servers holding shares is connected via private channels. Furthermore, it is sufficient if there is only one private channel between the client and one shareholder. In addition to improving practicability the presented proactive secret sharing solution, called EPSS, performs data aggregation to provide an efficient solution with respect to the communication complexity. Nevertheless, it still provides unconditional confidentiality for the data at rest and towards external attackers eavesdropping the communication channels.