Past student projects in 2022

Proposed projects page (archives): Spring-2022

Proposed projects page (archives): Fall-2022


Fall Semester

Optimizing Front-Running Protection

Michael Gysel

Report

Abstract

Auctions are a significant component of the global economy, with the three leading auction houses earning US$12.6B annually. Furthermore, online auctions account for an increasing percentage of auction sales at $5B in revenue annually. The most common auction forms are open-bid auctions and sealed-bid auctions. In open-bid auctions, bids are transparent and bidders actively compete against one another until the highest bidder is selected. In sealed-bid auctions however, all bidders submit secret bids during the bidding period. When the bidding period is complete, the bids are unsealed and the highest bidder is declared the winner. Arguably the greatest advantage of sealed-bid auctions results from the sealing of bids stifling active competition. This prevents collusion between bidders and incentivizes bidders to bid their valuation of the asset.

 

Dissecting IPFS and Swarm to demystify
distributed decentralized storage networks

Sixiao Xu

Report

Abstract

Decentralized networks have emerged as a promising solution for building secure and decentralized infrastructure for a variety of applications. Decentralized storage networks, in particular,
play a crucial role in hosting and distributing large amounts of data in a decentralized manner. One of the key challenges is developing effective incentive protocols to reward participating nodes for their contributions to the network, which is essential for maintaining the network’s health and functionality.

BFT Baxos: Robust and Efficient BFT Consensus using Random Backoff

Zhanbo Cui

Report

Abstract

Consensus protocol can ensure distributed system nodes agree on a unique value from their proposed values if all the nodes follow the protocol correctly. It has four properties: validity, integrity, agreement and termination. However, some nodes might not follow the protocol.
They might omit to send messages or send contradicting messages to other nodes. If a node behaves arbitrarily, it is called a byzantine node. BFT consensus protocol should allow all correct nodes to reach a common decision despite the byzantine fault.

Making the Inter-Planetary File System (IPFS) more reliable

Qiyuan Liang, Yuening Yang

Report

Abstract

One of the main benefits that distributed file systems have over traditional file systems that are stored on a single device or server is data redundancy. To achieve data redundancy, distributed file systems usually use replication. However, it is debatable whether replication is the best option in distributed settings. Researchers have proposed schemas that have some advantages over replication in certain aspects, for example, lowering the storage overhead, requiring less maintenance effort, and improving recovery capability. One of these schemas is Alpha Entanglement (AE) codes. In this project, we try to integrate IPFS, one of the most widely used distributed file systems nowadays, with AE codes. Users could only use replication in file reliability solutions provided by IPFS, which could be a large storage overhead and resource contention. The intention of this project is to offer one more solution on IPFS for file reliability. With AE codes integrated into IPFS, users could achieve a similar reliability level as replication with less storage and a slight compromise in the bandwidth overhead. Our solution builds on top of both IPFS and IPFS Cluster. Users are able to define and control their file reliability. The evaluation results
suggest a 2.7× storage saving, with a maximum 70% more bandwidth usage for file recovery, compared to replication.

 

D-voting Production Readiness

 Amine Benaziz, Albert Troussard

Report, Presentation (common)

Abstract

Electronic voting systems have the potential to improve the transparency and trustworthiness of elections by using blockchain technology to provide a secure record of all votes casted. DEDIS has been a pioneer in the e-voting field with its D-Voting project, an e-voting system based on the high-level blockchain components in the DELA project. In our work, we thoroughly implemented a series of automatic tests to fully evaluate the correctness, performance, robustness of the system in realistic scenarios and we were able to successfully improve its usability. While our work allowed the system to be overall more reliable, secure, and easy to use, a problem when the system is receiving numerous transactions in a short time has been discovered and important leads have been found towards its resolution . Overall, our work on the D-Voting system has helped to ensure that it is secure, usable, and reliable, and can soon be ready for use in high-stakes elections where trust is of the utmost importance.

D-Voting – Front-end development for an e-voting platform

Ahmed Elalamy , Ghita Tagemouati , Khadija Tagemouati

Report, Presentation (common)

Abstract

To address the growing importance of e-voting in our lives, the DEDIS lab has been working with students to create new solutions. The D-Voting project, which is based on the Dela blockchain, promises to provide a system that guarantees voter anonymity while being completely auditable and decentralized. A web application was previously created in order to facilitate the communication with the blockchain itself. As a result, the objective of this project’s fourth iteration is to enhance the front-end features that are already in place and add new ones.

D-Voting – Security audit

Chen Chang Lew

Security audit, Report, Presentation (common)

Abstract

D-voting project aims to launch and provide service to EPFL users in 2023. My main job is to conduct a security analysis for this open-source project. Conducting a security analysis in this project is important for several reasons. First and foremost, it helps protect the users of the project by identifying and addressing potential vulnerabilities that could be exploited by malicious actors. It also helps protect the reputation and integrity of the project by demonstrating a commitment to security and responsible development practices. Additionally, it can improve the overall quality and stability of the project by identifying and addressing potential issues
before they become widespread. Overall, conducting a security analysis on this project is a critical step in ensuring that the project is safe, reliable, and trustworthy for all users before we launch this service.

Front-running Protection on Ethereum

Shufan Wang

Report

Abstract

Nowadays, front-running attacks have been prevalent in many blockchains such as Ethereum, where a miner can inspect and order the transactions in a block with one’s preference to arbitrage. The profit of this attack, so-called MEV (Miner Extractable Value), has grown to a billion market.
Different approaches have been proposed to defend against front-running attacks. One promising approach is adding another consensus layer on the content of a block, which decentralizes the privilege of ordering transactions. However, this approach requires remarkable changes to the existing blockchains’ consensus layer. Another ideal solution employs a two-phase commit-and-reveal scheme, where miners have to order transactions in a blind manner. This solution introduces remarkable latency as it requires multiple rounds of interactions between the user and the blockchain.

Spring Semester

Anonymous Proof-of-Presence Groups for Messaging, Voting and Digital Currency

Robin Goumaz, Nico Hauser, Louis Bettens, Diego Boros, Maxime Zammit, Johann Plüss, Anders Hominal, Stefan Eric, Romain Birling, Zoé Marin, Ajkuna Seipi

Report

Abstract

Online communication plays a big role in everyday life: People communicate with other people using messengers and read content posted by others in social networks. In contrast to conventional conversations in the real world, in the digital world it is simply not possible to determine what person you are talking to without having met them in real world. A user id or display name cannot be authenticated and reliably linked to a real person. Even for verified profiles you have to rely on the service provider’s verification mechanism. And things are actually even worse: You do not even have the guarantee that the other party is a real person. In almost all online services there is a many-to-many relationship between online accounts and real people: Many real people can own none or a single online account and a single real person can own none to multiple online accounts.

D-Voting – Front-end development for an e-voting platform

Badr Larhdir, Capucine Berger

Report, Presentation

Abstract

To answer to the rising place that e-voting has taken in our lives, the DEDIS lab has been work- ing alongside students to develop new solutions. Based on the Dela blockchain, the D-Voting [2] project is an open platform that aims at providing a system guarantying anonymity of the votes, while being fully auditable and decentralized. As for the communication with the blockchain itself, a web application was developed to facilitate these interactions. Thus, as the third iteration of this project, the goal is to improve on the pre-existing functionalities and to implement new features in the front-end.

D-Voting – Engineering a Production-Ready System for an e-voting platform

Guanyu Zhang, IGOWA Giovanni

Report, Presentation

Abstract

Electronic voting system has drawn more and more attention in the recent years. DEDIS lab has initiated a distributed voting system based on DELA blockchain, called D-voting. After a couple of iterations, the D-voting system becomes a functional prototype ready for deployment. In our work, we first went thoroughly through the existing system to find critical steps. We then put in place a testing pipeline on top which our tests are executed. Afterwards we implemented a series of automatic tests to fully test the correctness, performance and robustness of the d-voting system in realistic scenarios. At the end, we investigated the errors occurring during the tests and proposed some modifications and other ways for further investigations in order to bring the system fully production ready.

Numerically-intensive deterministic smart contracts

Alexis Schlomer

Report

Seven years ago, Ethereum was launched and the world saw its first decen- tralized smart contract platform in operation. It held the promise of replac- ing what we now consider indispensable centralized intermediaries, such as banks, insurance companies and various governance structures, on a massive scale. However, there is still a significant gap between its ultimate goals and what is technically feasible today.

Vulnerability Assessment of Swiss Post E-Voting System

Vladyslav Zubkov

Report

For years, Swiss Post, Federal Chancellery, and independent researchers have been developing, reviewing, and creating legal frameworks for Switzerland’s E-Voting system for elections and referendums. The system follows two main goals – being convenient for voters and, at the same time, ensuring the secrecy and integrity of votes. The system, specification, and development processes were made public. Federal Chancellery defined and published a draft of the Ordinance on Electronic Voting to facilitate and organize the goals that the future E-Voting system should follow. Since then, independent researchers have provided additional input to the Swiss Post to make the system more and more secure.