Please note: This seminar will be given online.
Sujaya Maiyya
Computer Science Department
University of California, Santa Barbara
Modern applications produce over 2.5 exabytes (1018) of data every day and much of this data — including private and sensitive data — is stored on and maintained by third-party cloud providers. This outsourcing of data makes secure, yet efficient, data management more critical than ever. This talk outlines my efforts in developing efficient, fault-tolerant, and secure data management protocols.
Typically cloud enterprises replicate the data for fault-tolerance and shard (or partition) the data to store the shards on multiple servers for scalability. So far, data fault-tolerance (achieved using consensus protocols) and data scalability (supported by atomic commitment protocols) are viewed as independent and disjoint problems. My research work proposes an alternate perspective on this by unifying the seemingly disparate consensus and atomic commitment into a single framework called the Consensus & Commitment (C&C) framework. This unified approach facilitates developing new protocols that improve the performance, fault-tolerance, and security of data management systems. I will highlight the potential of the abstract C&C framework by deriving two concrete protocols across different contexts. First, I will propose a novel fault-tolerant transaction commitment protocol for shared and replicated data architectures. Second, I will present a replication protocol for the first fault-tolerant privacy preserving (oblivious) datastore that addresses some of the growing data privacy concerns.
Bio: Sujaya Maiyya is a PhD candidate at UC Santa Barbara, co-advised by Amr El Abbadi and Divy Agrawal. Her research interests lie in the intersection of data management, distributed systems, and data security. Sujaya is a recipient of IBM PhD Fellowship and Google PhD Fellowship (declined) awards for 2020–22, who is also recognized as an Outstanding Graduate Student (2019) for her commitment to the Computer Science Department at UCSB. She has spent her summers solving problems in distributed systems at companies such as Google and IBM Research. She is a core member of Diversity & Inclusion in Databases initiative and serves her community as a tutor for temporarily homeless children through Schools on Wheels.