Trading System for Confidential Market Execution

Founded 3 years before turning to Devox, the startup aimed to revolutionize the way market participants interacted, providing an anonymous platform for broker-dealers and market makers to improve trade execution quality.

Despite its promise, the platform ceased operations after a few years, but not before making a significant impact in the alternative trading space.

Our client approached Devox Software with a specific need: to develop a trading platform that allowed market makers and order-entry broker-dealers to interact anonymously. The key goal was to improve the execution quality of trades by facilitating real-time interactions between these two types of participants.

Given the sensitive nature of large institutional trades, anonymity was crucial to preventing adverse market impact, such as price slippage, caused by information leakage.

Anonymous Interaction

The project’s COO wanted to create a platform where market makers could send and receive anonymous trading signals, while order-entry broker-dealers could access market liquidity without revealing their identities or intentions. This interaction had to be seamless, secure, and fast, especially since high-frequency trading and large institutional orders were involved.

Execution Quality

Another core requirement was ensuring that the platform improved execution quality for broker-dealers. This meant building a system that provided:

• Access to multiple liquidity pools, allowing for price improvements and minimal slippage
• Advanced algorithms for trade matching and execution optimization.

Platform Architecture

Devox Software designed:

• Microservices-based architecture to enable different components of the trading system to interact independently yet cohesively. 
• .NET Core back-end was built to handle thousands of orders and interactions per second, ensuring low-latency communication between market makers and broker-dealers. 
• Kafka for real-time data streaming, which allowed market participants to receive and react to market signals almost instantly.
• Redis as an in-memory data store, optimizing the system for high-speed data access. This minimized lag in data retrieval, helping market makers and brokers see real-time updates to stock prices, orders, and liquidity pools.

Anonymous Interaction and Trade Matching

• Sophisticated masking and encryption algorithms were implemented to enable anonymous trading, which ensured that market makers and order-entry broker-dealers could interact without revealing sensitive information. 
• OAuth 2.0 and token-based authentication was implemented to make sure the platform is only accessed by authorized participants
• For trade matching, Devox built custom matching algorithms that aggregated liquidity from various market makers. These algorithms optimized trade execution by finding the best available price and executing trades without revealing the size or direction of the order to the broader market. This anonymity preserved the confidentiality of large trades, preventing market impact.

Front-end Design and User Experience

• React.js, ensuring that traders had access to real-time data with zero lag. We designed user interface in a way to allow broker-dealers to see liquidity from multiple market makers at a glance, with a customizable dashboard that showed key metrics like stock prices, IOIs, and order history.
• WebSockets ensured continuous communication between the back-end and front-end, updating the platform in real-time without page reloads.

Quality Assurance and Testing

Given the mission-critical nature of the platform, Devox conducted extensive quality assurance (QA) and testing:

• A combination of automated tests using Selenium and manual testing to ensure that the system could handle high trading volumes and complex interaction patterns.
• Load testing using JMeter, simulating thousands of simultaneous trades and interactions to ensure the platform remained stable under heavy load. 
• Security testing included vulnerability assessments to ensure that no sensitive data could be leaked during anonymous interactions.