free responsive website templates



Sikoba Research

We are working on cryptography, blockchain and distributed systems

About

Our purpose is to conduct fundamental and applied research in the areas of cryptography, blockchain and distributed systems. The goal is not just to write academic papers and file patents, but also to develop software, produce didactic material and participate in setting blockchain standards.

Incorporated in Luxembourg, Sikoba Research Sàrl is a sister company of Sikoba Ltd which is developing a peer-to-peer IOU platform based on blockchain technology. 

Research Topics

  • Verifiable Computing and Privacy-Preserving Smart Contracts -  We are developing a framework that allows to delegate execution of smart contracts away from the blockchain while keeping proof-of-correct execution on-chain, thereby preserving the fundamental principles of decentralised ledger technology.
  • Consensus Protocols - We are working on methods to optimise and speed up consensus protocols for consortium blockchains, in particular one-step consensus and self-pruning networks. We are also developing an approach for non-deterministic state machine replication.
  • Token economics - As the blockchain hype is receding, we need new approaches to how tokens are used on smart contract platforms. We are exploring models in which transaction prices are naturally linked to costs of cloud  computing and cloud storage.

Publications

A Primer on NIZK Proofs for Secure Computation - Stéphane Vincent - Dec 2018

Non-interactive zero-knowledge proofs (NIZKPs) are an essential element for providing security, confidentiality and privacy on public smart contract platforms. After introducing the necessary background on circuit representation, a description of a C-to-arithmetic-circuit compiler is provided. Finally, two NIZP systems based on different cryptographic assumptions are presented.

Exploring Pairing-Based Cryptography - Stéphane Vincent - Dec 2018

One of the key cryptographic primitives behind various constructions, including privacy-preserving authentication, short signature schemes, and zero-knowledge proofs is the bilinear mapping technique that uses pairings over elliptic curves. Initially used in cryptography to break the discrete logarithm problem in a group of points of some elliptic curves, pairings are now considered to be one of the most suitable mathematical tools to design secure and efficient cryptographic protocols. 

State of the Art in Verifiable Computation - Dmitry Khovratovich - Dec 2018

This paper presents a detailed overview of current approaches used: Interactive Proofs, Homomorphic public-key cryptosystems, Discrete logarithm problem, MPC derivative and Incrementally Verifiable Computation. The properties of existing implementations are discussed and their performances compared based on available benchmarks.

Verifiable Databases - Dmitry Khovratovich - Dec 2018

An overview of current approaches for verifiable database interactions. When only queries are involved, the solutions only need to prove authenticity. With updates, computations integrity must be proven.

One-Step Consensus in Weakly Byzantine Environments Aleksander Kampa - revised and expanded version March 2019 (original version was Dec 2018)

Under some conditions, binary asynchronous Byzantine consensus can be reached in a single communication step. This paper show how these conditions can be relaxed when only a subset of faulty nodes is Byzantine.

J-R1CS, a JSON Lines format for R1CS - Guillaume Drevon - February 2019

R1CS (rank-1 constraint systems) define a set of bi-linear equations which serve as constraints suitable for ZK proofs. This paper presents J-R1CS, a simple and lightweight JSON Lines format dedicated to R1CS.

Work -in-progress:


Fast / Timeout-Resistant / Network-Pruning Consensus in Weakly Byzantine Environments - Aleksander Kampa - March 2019 (early draft, feedback requested)
By spawning multiple Single-Sender consensus processes, and introducing the concept ofproofto prevent premature timeout messages form Byzantine nodes, the MPC protocol achieves efficient consensus that is very fast under favourable circumstances..

Software

Mobirise

Our verifiable computation software project is called Isekai. The first module that is being developed parses C programs and outputs the arithmetic and/or boolean circuit representing of the expression equivalent to the input program. Isekai uses libclang to parse the C program, so most of the preprocessor (including the includes) is available. Isekai is written using the Crystal programming language allowing for strong type safety and is compiled to a native executable, ensuring maximum efficiency in parsing.

Main Sponsor

Fantom Foundation

Based in South Korea, Fantom is the world’s first DAG-based smart contract platform that solves the scalability issue that has long plagued existing public distributed ledger technology. Built on Fantom’s OPERA Chain, the Fantom platform will replace existing payment methods and supply chain management infrastructures using a dApp that will handle hundreds of thousands of transactions per second across all industries, including food, telecommunications, finance, electricity, electronics, real estate, and autonomous vehicles, saving costs and providing transparency.
Mobirise

Partner with Us

Partnership and Internship Opportunities

We are looking to partner with researchers who work on similar topics of interest.

Types of collaboration can include:
+ commissioned research reports or papers
+ co-authorship of research papers
+ ongoing research collaboration on a specific topic
+ research assistance

On the software side, we are looking to cooperate twith talented developers who interested in topics such as programming language design and parsing, cryptography and distributed systems. Our language of choice is Crystal.

We are also looking for interns who are interested in researching and/or doing development work on topics related to distributed computing, blockchain and cryptography.

Address

Sikoba Research is a project of Sikoba Ltd for more information visit sikoba.com

Contact

Email: info@sikoba.com