Universal Derivative Trading Platform
A leading US-based technology and financial engineering company was looking to develop a cloud-based trading platform geared toward the U.S. financial market.
The system is a clustered, highly load-resistant solution that uses Amazon Elastic Compute Cloud (Amazon EC2), and intricate mathematical optimization algorithms to solve advanced tasks. It is fully integrated with the financial market data providers.
The system’s functionality supports full-cycle trading across multiple asset classes, including derivatives, such as US treasury bonds, IRS (interest rate swaps), and more. The general workflow for the system is as follows:
- A member trader can submit trades, receive information on the trades that have been submitted, and confirm that the submission is accurate
- The system makes the closing prices for a corresponding product available to the member firms 15 minutes after the market is closed.
- All confirmed submissions are declared final by the system within 30 minutes after the closing prices are made available.
- The system then notifies the traders that the deals have been closed, and starts running the appropriate algorithm.
- The trading results are displayed to the trader, and sent to external services.
The trading platform also recommends potentially profitable deals, and provides a flexible API though which it can be managed.
INSART was chosen to implement the project out of several candidate companies due to the technical skills and previous experience of its proposed project team.
The project team we put together started by actively gaining additional knowledge in the subject field of the project.
The number of the INSART experts deployed on the project varied from 3 to 5 people during the different phases of the implementation effort. The project team included a project manager, one to four software engineers, one business analyst, one QA engineer, and one mathematical algorithms’ developer and analyst.
We were actively involved throughout the development cycle, from the definition of the system requirements, to the system’s deployment on the client’s side. This also included the application’s deployment in the Amazon cloud, and direct communication with the external services associated with the system’s functioning.
The project posed several challenges, the more significant ones being the following:
- The project required a great deal of highly specific knowledge associated with off-the-counter trading. The team had to be trained on the specifics of the business domain within a very short period of time in order to be able to perform the development efficiently.
- The solution had a complex multilayer software architecture that was supposed to be integrated with market data providers and order confirmation services. It was also required that the system have good usability for both traders and the client’s operators. We managed to suggest a robust service-oriented architecture that can be re-used to support various financial instruments.
- The performance of the optimization algorithm was to take not more than 5 minutes after the market was closed. Our math algorithm developer was able to suggest and develop a highly optimal solution using a R framework and distributed computing system.
- It was quite difficult to gain sufficient knowledge about off-the-counter trading. This necessitated active communication with traders from several investment banks’ trading desks, who did not have too much free time. We had to establish a very efficient communication process to be able to receive the information we required.
INSART succeeded in meeting all the above challenges, and delivered, on time and on budget, a system that was in full compliance with the client’s requirements.
The INSART team has been using the following technologies to implement and support the project:
- Java/JEE (JBoss, PrimeFaces, Hibernate, etc.).
- Amazon EC2 for hosting, clustering and performance optimization.
- C# tool for Excel for operational application.
- Numerical computing environment based on R framework.
- Event System Esper.
- Web server Nginx.
- Integration of 3rd-party web services.
INSART delivered to the client a fully functional solution that has been truly instrumental for the client’s customer acquisition processes. They have been able to broaden both their customer base, and their range of the traded asset classes.
INSART was further engaged to develop new versions of the application, and provide technical and operational support of the system.