Distributed Key Systems: Enhancing Security, Fault Tolerance and Disaster Recovery in Cloud Computing

Cloud computing has increased the need for sound and sophisticated security strategies. Although encryption is the primary mechanism to provide security in data transmission, cloud computing offers both advantages and disadvantages to enhancing the encryption process. Encryption algorithms currently available are robust, but are well known and require a “key” to make a session unique. Therefore, if the key is compromised the encryption can be unlocked no matter how robust the algorithm. In the classical model, a node in a network contains the entire key, and if compromised, then potentially the entire cloud’s security could be breached. Also, if the key were destroyed then all the data it protects could be lost, because the data couldn’t be unencrypted. Therefore, splitting the key and storing portions on multiple nodes within a cloud would make it more difficult to compromise or lose the key. Further, the key could be striped (like a RAID, Redundant Array of Independent Disks) so if distributed across multiple nodes some could fail and the key could still be recovered. This methodology could create greater security, while providing greater fault tolerance by allowing retrieval of the entire key even if some nodes were compromised or destroyed.