60 likes | 147 Vues
This research explores how optimization affects user privacy in distributed database systems. It addresses balancing privacy and performance, revealing sensitive information, network traffic implications, and empowering users to specify privacy preferences. The study proposes considering privacy as an optimization criterion in query processing to enhance data security. Implementation is done within PostgreSQL, utilizing constraint/preference data structures for optimization and developing a parser for extracting privacy constraints from SQL queries.
E N D
Privacy in Distributed Database Systems Adam J. Lee, Panos K. Chrysanthis, Nicholas L. Farnan adamlee@cs.pitt.edu
Distributed Query Processing SELECT * FROM Plants, Supplies, Polluted_WatersWHERE Supplies.type = "solvent", AND Supplies.name = Polluted_Waters.pollutant, AND Polluted_Waters.location = Plants.location, AND Plant.id = Supplies.plant_id; Alice Querier Inventory Facilities Pollution Watch Trusted Untrusted
How Does Optimization Affect Querier Privacy? SELECT * FROM Plants, Supplies, Polluted_WatersWHERE Supplies.type = "solvent", AND Supplies.name = Polluted_Waters.pollutant, AND Polluted_Waters.location = Plants.location, AND Plant.id = Supplies.plant_id; Reveals sensitive information to ManuCo Reveals sensitive information to Pollution Watch Results in a large amount of network traffic Strikes a balance between privacy and performance
SELECT * FROM Plants, Supplies, Polluted_WatersWHERE Supplies.type = "solvent", AND Supplies.name = Polluted_Waters.pollutant, AND Polluted_Waters.location = Plants.location, AND Plant.id = Supplies.plant_id; The 10,000 Mile View… Make sure all operations involving these conditions are evaluated by a trusted server! Empower users to… • Identify sensitive intensional regions within their queries • Specify constraints/preferences over how these regions are handled Enable the system to… • Consider privacy as an optimization criterion • Construct query plans respecting these constraints and preferences
So where are we? PROOFS … Implementation within PostgreSQL
What would you do? 3. Generate constraint/preference data structures needed for optimization 1. Learn flex and bison 2. Develop a parser for extracting privacy constraints from SQL queriers 4. Integrate with our modified version of PostgreSQL