simukhop «at» ucsc·edu
Sinjoni is a third year PhD student working under Professor Darrell Long in the Department of Computer Science at the University of California, Santa Cruz.
Archival storage, storage security, information retrieval, cloud storage architecture, data privacy in cloud, cryptography and secure data in Internet-of-Things.
To avoid the complexities of key management and constant updates due to encryption algorithm deprecation, authentication is preferred over encryption to securely store archival data. Secret-Splitting is a method that not only improves data reliability, but also reduces risks of insider threats. SSRC has had projects like POTSHARDS in the past that has been used to reconstruct a secret-split archival datastore. We are currently working on improved solutions for reconstruction of secret-split datastores. The Secret-Split Secure Hash Method (S3HA) is used to quickly reduce reconstruction space of secret-split datastore required to identify sibling shares. It uses a smaller subset of bits (hints) to identify shares belonging to the same object. Duplicate computations are avoided by testing for precomputed pairs against a bit vector.
Despite the multitude of advantages that the cloud has over on-premise storage, when it comes to access control and privacy of IoT data storage using a single cloud provider has major disadvantages. The storage provider has complete control over secure computations on the data which means they can use the data for their own purposes. Additionally, the user depends solely on the provider to access all the data where the single point of access has adverse impacts. Any failure on the provider's side could result in complete loss of data and data could easily be compromised if the storage provider is broken into.
We are working on a distributed storage model that leverages secret-sharing properties to securely store data from IoT devices.