This project, named GeoENS, augments existing DNS infrastructure to support low latency DNS lookups for edge applications. There are many use cases for traditional GeoDNS systems, like Amazon's Route53, in the centralized web. Use cases include proximity-based load balancing and serving content specific to the geographic location of the query. Unfortunately existing blockchain-based DNS solutions (namely ENS) does not have a mechanism for geo-localized query resolution, or more technically, geographic split horizon DNS. This work enables geographic awareness for blockchain-based DNS. GeoENS is a smart contract specification for decentralized location services built on the Ethereum Blockchain. The specification including a reference implementation is open source, EIP (Ethereum Improvement Proposal) 2390. Preliminary results were accepted to USENIX HotEdge'20 workshop and full results are underway. EIP - Source Code
In many decentralized applications, messages are signed by a
single key whose owner is considered the single party
responsible for the current operation, regardless of if
that single party is an individual or an organization.
An alternative approach is to split the key into multiple pieces
and distribute the signing process between multiple parties.
This process can be done using secure multiparty computation (MPC)
and is desirable in situations where a single party may
not be responsible for signing.
This project interfaces with Unbound Tech's secure multiparty computation (MPC) based ECDSA Signature Library. An ECDSA signature is jointly computed between two parties and stored in a smart contract, deployed on a local blockchain. The chain consists of geth running Proof-of-Authority consensus algorithm. Source Code -
This project was a computer program similar to Pandora that would give recommendations for music. Given a song, the program would select 5 songs from it’s library determined to be the most similar. The program used DSP techniques such as spectral centroid tracking, spectral periodicity, and analysis of energy in spectral bands to extract features from each of the songs in its library. After feature extraction, the program then ran the K-NN machine learning algorithm to determine which songs were most similar. The project won first place in the individual undergrad category at Engineering Expo with over 5000 in attendance. Source Code
A decentralized application which act as a living will and testament. The DApp, consisting of a smart contract, stores a users cryptocurrency. In the event of death, their funds are distributed amongst their planned benefactors, without the need for lawyers, mediation, or arbitration. Source Code
A fully electronic system that automates the process of brewing beer. Used Raspberry Pi to control electronic heating element via PID algorithm. Temperature set-point adjusted via Apache hosted website. Source Code - Video