Sulphur
A High Frequency Prediction Market
Demo
About
The problem it solves #Scaling Prediction Markets with High-Performance Blockchain Architecture ## Current State of Prediction Markets Prediction markets on chains like Polygon (~7k TPS) and Solana (~65k TPS) fall short compared to HFT systems processing millions of orders per second. Scalability issues in transaction throughput and order matching hinder growth. ##Vision for Scalable Prediction Markets Inspired by [Vitalik Buterin’s insights on information finance](https://vitalik.eth.limo/general/2024/11/09/infofinance.html), we recognize that existing prediction markets lack the infrastructure to support a vast number of simultaneous markets and order books. Our vision is to enable platforms like X to allow hundreds of thousands of users to create new prediction markets effortlessly.Potential use cases include: - **Mass Market Creation:** Enabling platforms like [XPredict](https://ethglobal.com/showcase/xpredict-2g4st) to support 100k users generating new prediction markets daily. - **Academic Integration:** Facilitating hundreds to thousands of scientific reviews by academics each day. - **Social Media Synergy:** Utilizing prediction markets as dynamic social media tools and information aggregators. ##Our Solution: Custom Layer 1 and Optimized VM To achieve scalable prediction markets, we propose a dual-layer approach: 1. Custom Layer 1 (App Chain): Developing a dedicated blockchain optimized for prediction market operations, ensuring high throughput and low latency. 2. Custom-Tuned Virtual Machine (VM): Utilizing HyperSDK, we have developed a specialized VM with embedded order book logic Our platform ensures flexibility and scalability through: - Multi-Chain Deployment: Deployed on Polygon and Base testnets, with the Okto Abstraction layer enabling seamless migration to chains like Avalanche L1. - Performance Optimization: VM-embedded order matching reduces overhead and boosts TPS. Performance testing will refine this as HyperSDK matures.
How it's made
Challenges we ran into # Okto The Avalanche and Base testnets are not supported on Okto. We ran into a lot of issues integrating it. But it works now. Yay. # HyperSDK Very limited documentation. We had to go deep through the code implementations and design an OrderBook and matching algorithms inside a VM. #Etna TestNets Recent updates to etna testnets dont have support to deploy my custom vm using HyperSDK. So a major part of product could not be delpoyed #Network Issues working on blockchian requires me to be onchain