Non-interactive Cryptographic Timestamping Based on Verifiable Delay Functions

Abstract

We present the first treatment of non-interactive publicly-verifiable timestamping schemes in the Universal Composability framework. Similar to a simple construction by Mahmoody et al., we use non-parallelizable computational work that relates to elapsed time to avoid previous impossibility results on non-interactive timestamping. We extend these ideas to the UC-framework and show how to model verifiable delay functions (VDF) related to a global clock, and non-interactive timestamping, in the UC-framework. Furthermore, we present new constructions that are substantial improvements over Mahmoody et al.'s construction, such that any forged timestamps by the adversary are now limited to within a certain time-window that depends only on its ratio to compute VDFs more quickly and the time-window of corruption. Finally, we discuss natural applications for our construction in decentralized protocols.

RandRunner: Distributed Randomness from Trapdoor VDFs with Strong Uniqueness

“…VDFs were first introduced by Boneh et al [6] in 2018, and have since received increased attention from other researchers (see, e.g., [6], [46], [39], [7], [20], [32], [44], [25], [24], [34]). As introduced by Boneh et al [6], [7], a VDF is a function f : X → Y which maps every input x ∈ X to an unique output y ∈ Y. Computing the VDF is sequential in the sense that it takes a prescribed amount of time, whether or not it is executed on multiple processors.…”

Timed Signatures and Zero-Knowledge Proofs—Timestamping in the Blockchain Era—

“…In fact, there are existing commercial solutions, e.g., Guardtime 4 , that use this idea to offer a blockchain-based timestampting system. Following this research line, very recently [29] presented a treatment of non-interactive timestamping schemes in the UC-model. The construction provided in [29] is based on proofs of sequential work such as VDF's [9].…”

“…Following this research line, very recently [29] presented a treatment of non-interactive timestamping schemes in the UC-model. The construction provided in [29] is based on proofs of sequential work such as VDF's [9]. However, as the authors stated in [29], the construction allows the adversary to pretend that a record was timestamped later than it actually was (i.e., it allows postdating attack).…”

“…The construction provided in [29] is based on proofs of sequential work such as VDF's [9]. However, as the authors stated in [29], the construction allows the adversary to pretend that a record was timestamped later than it actually was (i.e., it allows postdating attack). Also, even if the work of Landerreche et al assumes the existence of a global clock, the timestamping service provides only ordering of events 5 .…”

“…Concretely, we devise a formal model of protocol execution for timestamping cryptographic primitives with respect to a global clock that parties have access to. We use the term timed, as in timed (digital) signatures to distinguish timestamping with respect to such a global clock from the guarantee offered by existing timestamping schemes [25,5,29], which only establishes causality of events-i.e., which of the hash-chained document was processed first-but does not necessarily link it to a global clock. We stress that although for simplicity our treatment assumes ideal access to a global clock-which is captured as in [4] by a global clock functionality, it trivially extends to allow for parties having bounded-drift view of the clock [26]-i.e.…”