Privacy preserving cryptographic protocols

This weeks read was about privacy preserving cryptographic protocols.

The common definition of privacy in the cryptographic community limits the information that is leaked by the distributed computation to be the information that can be learned from the designated output of the computation (Benny Pinkas, HP Labs).

This chapter discusses differing protocol frameworks used to achieve this privacy. In Acquisti el al’s:Digital Privacy their definition is a computational function of outputs that are distributed among different participants during online collaboration.

The chapter notes a couple of platforms are in place today that have the goal to provide a privacy-preserving protocol for any possible function:
Secure Multiparty Computation (SMC)
Secure Function Evaluation (SFE)
and goes on to say that it may seem like an impossible task, but general results show that any function that is computable in polynomial time can be computed with polynomial communication.

The chapter goes on to discuss how privacy preserving cryptographic protocols can be applied to differing situations:
Database querying
Distributed voting
Bidding and auctions
Data mining

The Data mining or data warehousing kind of caught my attention because I had to write a white paper on it when it first appeared on the IT table. It was suppose to be the way of the future for storing and locating data quickly and easily. At that time privacy issues were not looked at closely as systems were not as integrated as they are today. Data mining and warehousing fell into the privacy conundrum with the advancements in technology and the widely integrated system structure in place today. It is interesting to see how this is handled by SMC and SFE to provide privacy security.
A couple of approaches were mentioned: Sanitizing Data before making it available and the use of technologies mentioned in the chapter, benchmarking and forecasting, contract negotiations, and rational selfish participants along with the introduction by Lindel-Pinkas method where two parties build a decision tree without either party learning anything about the other.