VaultDB: A Private Data Federation

For many people seeking actionable insights from data, their first major obstacle is getting access to relevant datasets. Despite the abundance of information we collect on practically every domain of life, much of it is fractured among many private data stores. For example, electronic health records on a given patient are often partitioned among multiple hospitals. Querying fractured datasets is often challenging owing to regulatory requirements and privacy concerns. A private data federation (PDF) queries the union of the private records of multiple autonomous data stores such that no one learns about the data of its peers. VaultDB, our PDF prototype, translates SQL queries into secure multi-party computation protocols and orchestrates their execution among the data providers. These cryptographic protocols incur substantial overhead for PDF queries with runtimes typically three or more orders of magnitude slower than that of an insecure execution of the same query. We are researching a query optimizer for VaultDB that leverages properties of the relational model to enable PDF queries to run efficiently. Papers: [pdf], [pdf], Code: [github]

BigDAWG: A Federator for Heterogeneous Data Models

BigDAWG explores a new view of data federations to address the growing need for cross-database querying over heterogeneous data stores. This need is fueled by the proliferation of storage engines and data models - such as arrays, graphs, and text engines - each usually having its own query language and storage semantics. We are studying how to integrate and optimize queries that span many disparate data models through a single query interface. Papers: [pdf], [pdf], Code: [github]

Systems for Emerging Data Scientists and Data Enthusiasts

We are investigating novel data models and query frameworks for non-traditional data analysts. We have researched this in two domains: personal data-driven decision making and querying for scientists conducting research. In the latter, our focus is on principles and techniques for making the scientific method a first-class citizen in the database. We designed these frameworks based on interviews with individuals who are presently underserved by relational databases who also have an abundance of data. We have synthesized our initial findings from these studies into two vision papers. Papers: [pdf] [pdf]

Johes Bater, Yongjoo Park, Xi He, Xiao Wang, and Jennie Rogers. SAQE: Practical Privacy-preserving Approximate Query Processing for Data Federations. In the Proceedings of the VLDB Endowment, 2020. [pdf]

Jennie Rogers, Johes Bater, Xi He, Ashwin Machanavajjhala, Madhav Suresh, and Xiao Wang. "Privacy Changes Everything," in POLY Workshop at VLDB, pages 96-111, 2019. [pdf]

• Presented at HPTS 2019: [slides]

Johes Bater, Xi He, Will Ehrich, Ashwin Machanavajjhala, and Jennie Rogers. "Shrinkwrap: Efficient SQL Query Processing in Differentially Private Data Federations," PVLDB, 12(3), 2018. [pdf]

Johes Bater, Greg Elliott, Craig Eggen, Satyender Goel, Abel Kho, and Jennie Rogers, "SMCQL: Secure Querying for Federated Databases," in PVLDB, 10(6), pages 673-684, 2017. [pdf] [code] [slides]

Jennie Duggan and Michael Brodie, "Hephaestus: Data Reuse for Accelerating Scientific Discovery," in Proceedings of CIDR, 2015. [pdf] [slides]

• Extended version: [pdf].

Jennie Duggan, Aaron Elmore, Michael Stonebraker, Magdalena Balazinska, Bill Howe, Jeremy Kepner,et al., "The BigDAWG Polystore System," in Sigmod Record, 44(3), 2015. [pdf]

Jennie Duggan and Michael Stonebraker, "Incremental Elasticity for Array Databases," in SIGMOD 2014. [pdf]

Jennie Duggan, "The Case for Personal Data-Driven Decision Making," in VLDB, 7(11), pages 943-946, 2014. [pdf]

Jennie Duggan, Ugur Cetintemel, Olga Papaemmanouil, Eli Upfal, "Performance Prediction for Concurrent Database Workloads," in SIGMOD, 2011. [pdf]

Philippe Cudre-Mauroux, Hideaki Kimura, Kian-Tat Lim, Jennie Rogers, Roman Simakov, Emad Soroush, et al., "A Demonstration of SciDB: A Science-Oriented DBMS," in VLDB, pages 1534-1537, 2009. [pdf]

---

---

---

---