Prof. Ashish Choudhury shares his thoughts with Naviiina on authoring a book on Secure Multi-Party Computation (MPC) in a time when big tech like Google, Facebook, and Microsoft have acknowledged the significance of MPC as a promising futuristic privacy-enhancing technology.
Secure Multi-Party Computation (MPC) is a fundamental concept, both in cryptography as well as Distributed Computing. On a very high level, a MPC protocol allows a set of mutually-distrusting parties with their private inputs to jointly perform any computation on their inputs by keeping the respective inputs as private as possible. Examples of such computations include but are not limited to privacy-preserving data mining, secure e-auction, private set-intersection, privacy-preserving machine learning, etc. One may imagine a MPC protocol emulating the role of an imaginary, centralized trusted third party (TTP), which collects the inputs of the parties, performs the desired computation and publishes the result. Therefore, any task which can be carried out securely in the presence of a centralized TTP can be performed by decentralizing the role of the TTP by running a MPC protocol among the parties themselves. Due to its powerful abstraction, the MPC problem has been widely studied both by the cryptography community as well as the distributed computing community over the past four decades.
Ever since its inception, MPC is one of the widely studied topics in secure distributed computing. There are two major categories of MPC protocols, the first category is that of MPC protocols against passive corruption while the other category has maliciously-secure MPC protocols, while passive corruption deals with a benign adversary which simply eavesdrops on the state of the corrupt parties during the protocol execution, malicious corruptions deal with a more powerful adversary, who can control and force the corrupt parties to behave in any arbitrary fashion during the protocol execution.
Even though the passive corruption model may seem very weak, achieving security against such a benign form of adversary turns out to be non-trivial and demands sophisticated and highly advanced techniques. Moreover, very often protocols secure against semi-honest corruption, serve as the basis for protocols secure against a malicious adversary. We strongly believe that the topic of passively secure MPC protocols in itself is a very vast and important topic to be covered in a single course. The aim of this book is to focus only on the passive corruption model and present all the seminal possibility and feasibility results in this model ever since the inception of MPC, with formal security proofs. We present protocols both against computationally bounded as well as computationally-unbounded adversaries.
Detailed security proofs are provided for the seminal protocols. For a better understanding of the underlying concepts, pictorial illustrations are provided and several examples have been worked out.
We also present state-of-the-art efficiency improvement techniques in the domain of passively-secure MPC protocols. Our intention is to provide the readers with a detailed and comprehensive explanation of all the important concepts and techniques in this domain with an aim to unfold the evolution of this topic from its inception to till date.
The book is readable by General Audience Too
The topic covered in the book (MPC) is a very vast topic. Till a few years back, MPC was used to be considered a theoretical topic, but now people are deploying MPC as a privacy-enhancing technology. Big technology giants especially Google, Facebook and Microsoft have acknowledged the significance of MPC as a promising futuristic privacy-enhancing technology. Even though the topic is so important, unfortunately, there is no single book where one can find all the classical algorithms/protocols in this domain. More importantly, the security proofs of MPC protocols, which constitute a major bottleneck in understanding cryptographic protocols are very often not unified and difficult to understand. To learn these protocols, one has to explore various research papers, which is a very difficult task for a beginner since they find it very difficult to read and understand these research papers. In fact, when I was a Ph.D. student, I faced the same ordeal. Given this background, we aimed to make the book available to the general audience. The language and delivery of the material in the book are such that anyone who has a basic understanding of algorithms/data structure/discrete mathematics can easily understand the contents.
**There is a companion MOOC freely available on NPTEL based on the contents of the book – https://nptel.ac.in/courses/106108229
**These video lectures are very useful when viewed after reading the contents of the book.