The authors present a comprehensive study of efficient protocols and techniques for secure two-party computation - both general constructions that can be used to securely compute any functionality, and protocols for specific problems of interest. The book focuses on techniques for constructing efficient protocols and proving them secure. In addition, the authors study different definitional paradigms and compare the efficiency of protocols achieved under these different definitions.

The book opens with a general introduction to secure computation and then presents definitions of security for a number of different adversary models and definitional paradigms. In the second part, the book shows how any functionality can be securely computed in an efficient way in the presence of semi-honest, malicious and covert adversaries. These general constructions provide a basis for understanding the feasibility of secure computation, and they are a good introduction to design paradigms and proof techniques for efficient protocols. In the final part, the book presents specific constructions of importance. The authors begin with an in-depth study of sigma protocols and zero knowledge, focusing on secure computation, and they then provide a comprehensive study of the fundamental oblivious transfer function. Starting from protocols that achieve privacy only, they show highly efficient constructions that achieve security in the presence of malicious adversaries for both a single and multiple batch executions. Oblivious pseudorandom function evaluation is then presented as an immediate application of oblivious transfer. Finally, the book concludes with two examples of high-level protocol problems that demonstrate how specific properties of a problem can be exploited to gain high efficiency: securely computing the kth-ranked element, and secure database and text search.

This book is essential for practitionersand researchers in the field of secure protocols, particularly those with a focus on efficiency, and for researchers in the area of privacy-preserving data mining. This book can also be used as a textbook for an advanced course on secure protocols.

Zusammenfassung
In the setting of multiparty computation, sets of two or more parties with p- vate inputs wish to jointly compute some (predetermined) function of their inputs. The computation should be such that the outputs received by the parties are correctly distributed, and furthermore, that the privacy of each party's input is preserved as much as possible, even in the presence of - versarial behavior. This encompasses any distributed computing task and includes computations as simple as coin-tossing and broadcast, and as c- plex as electronic voting, electronic auctions, electronic cash schemes and anonymous transactions. The feasibility (and infeasibility) of multiparty c- putation has been extensively studied, resulting in a rather comprehensive understanding of what can and cannot be securely computed, and under what assumptions. The theory of cryptography in general, and secure multiparty computation in particular, is rich and elegant. Indeed, the mere fact that it is possible to actually achieve the aforementioned task is both surprising and intriguing.

Inhalt
and Definitions.- Definitions.- General Constructions.- Semi-honest Adversaries.- Malicious Adversaries.- Covert Adversaries.- Specific Constructions.- Sigma Protocols and Efficient Zero-Knowledge1.- Oblivious Transfer and Applications.- The kth-Ranked Element.- Search Problems.