Spartan Space Notes - Synthetix & Pyth Network - March 14 2024

Pyth Network Oracles & Governance: Bringing data onchain and powering Synthetix Perps

Introduction

• Joined by Jayant CTO of Douro Labs, a company working on Pyth, a decentralized high frequency low-latency oracle solution 

• Jayant has been interested in crypto since 2014, when he created a trading bot when the market was less competitive. 

A market data vertical for DeFi applications.

• Jayant explains that Pyth is a first-party Oracle that provides market data from licensed data providers, ensuring sustainability for both data providers and users.

• Jayant highlights how Pyth differentiates itself from other Oracle services by being a first-party Oracle, allowing data providers to retain ownership of their data and license it for distribution.

• Pyth provides real-time US equity prices through its network of data providers, including big traditional finance firms like Two Sigma and Jane Street.

• Pyth is designed to work across multiple blockchains, with price data available on over 50 blockchains in real-time, every 400 milliseconds.

• Pyth’s differentiating factor is a “pull-based” oracle versus a “push-based” which helps further reduce latency. 

Data aggregation and price feed distribution in blockchain.

• Data providers stream quotes to software, which sends prices to Pythnet chain.

• Jayant explains how data providers stream prices to Pythnet, where a program aggregates the data to prevent manipulation.

• The system uses Wormhole to bridge the price updates to other blockchains [from Pythnet], and contracts on each blockchain can verify the messages.

Decentralized data providers and security measures.

• Jayant explains that prospective data providers for the network must have first-party data and go through an onboarding process to be permissioned to publish prices for specific symbols.

• The number of high-quality data providers has increased over time, with over 100 data providers now on the list, including major crypto exchanges and trading firms.

• Jayant: Data providers are permissioned, reducing risk of malicious attacks.

• Reputation of data providers guards against exploits, making it less likely for individuals to collude.

• Having many data providers also helps alleviate any issue with a single provider

• Jayant doubts other permissionless systems can work due to security concerns

• Other oracle's potential slashing system may not be feasible due to economic incentives and data security concerns.

Integrating Oracles into DeFi protocols with latency considerations.

• Jayant: Decentralized Oracle systems are unlikely to be faster than high-frequency traders' optimized pipelines.

• Built-in delay is a natural part of the trading process, with prices and fills occurring within seconds of each other.

• Oracle data can have confidence intervals, which are used by protocols to make decisions based on the uncertainty of the data.

• Jayant suggests that protocols can use these confidence intervals to tolerate latency in certain situations, such as borrowing and lending.

Decentralized governance and price uncertainty in a decentralized protocol.

• Jayant explains the purpose of confidence intervals in market data, which represent uncertainty ranges for asset prices and can be used to incorporate dispersion information into downstream usages.

• There is an additional layer of risk analysis provided by confidence intervals for protocols ingesting and using this data.

• The governance model of Pyth is controlled by token holders through a decentralized governance system.

• Jayant discusses the challenges of decentralized governance and how Pyth Gov system is based on the Synthetix’s model, with token holders electing members to two councils: Pythian Council and Price Feed Council, each with different powers.

Decentralized governance models

• Community engagement and representation in direct tokenholder voting is limited, but representative council system could improve it.

• Jayant: Ideal governance system is no governance, but practical in some cases.

• Representative councils improves community involvement and shares insights on future features and improvements.

• Jayant discusses the company's focus on investing in the features of their Oracle product to maintain its lead in latency, reliability, and performance.

Synthetix Perps and Pyth Network solutions.

• Pyth's speed and efficiency in listing and dropping new tokens impresses, with dependence on smaller Oracle services decreasing.

• Matt discusses the evolution of Synthetix governance and the importance of Pyth in the space.