PHILIPPINE STOCK EXCHANGE – Case Studies
Philippine Stock Exchange, Inc relies on PostgreSQL as a backend of their new trading platform
THE BACKGROUND & CHALLENGE
The Philippine Stock Exchange (PSE) is one of the oldest stock exchanges in Asia, having operated continuously since the establishment of the Manila Stock Exchange in 1927. PSE selected NASDAQ as its new service provider for the Exchange’s trading engine in July 2014. NASDAQ deployed its X-Stream Trading System (XTS) to the PSE to replace the NSC trading system provided by NYSE Euronext Technologies SAS. The new trading system, called the PSETrade XTS, was being developed with enhanced trading capacity and increased risk management parameters in line with PSE’s overall strategy of introducing more products and services to the market. The backend RDBMS for PSETrade XTS was PostgreSQL, which was relatively new for PSE. With this situation, it was imperative for PSE to get their team trained on the new trading platform to make ensure efficient handling of daily chores for the production server.
THE SOLUTION DELIVERY
After examining PSE’s business situation and requirement, Ashnik and its local partner IT Group, commenced the process of implementation with a solution-approach. The teams studied the application’s design architecture and architected PostgreSQL database to optimally power and support the application.
Key contributions to the PostgreSQL setup for NASDAQ’s X-stream Trading System Platform:
Ashnik developed a robust, resilient and highly available solution using EDB Postgres’ Failover Manager and Backup Recovery Tool. Some of the key areas addressed were:
- Mechanism for quick response to read request – Incorporated a crucial mechanism to enable quick response to read requests from caching service within the application. The application’s high TPS (transactions per second) requirements were handled by caching all the data on the application cache.
- Database Tuning –
– Tuning as per nature of application: The application using PostgreSQL as backend does a bulk read (at start of each day). Then at end of the day, data is flushed to a persistent database. Database was tuned to optimize these bulk operations
– Tuning as per setup and environment: The database is co-hosted with other application components which are memory-intensive. Database was tuned in accordance to not do aggressive caching and deprive the application of memory
- Backup-Recovery & DR – Gave actionable inputs to PSE on how to design its Backup-Recovery and DR strategy using EDB Postgres’ Backup Recovery Tool
- Security – Secured the Postgres database by restricting the connections with Host Based Authentication configuration of PostgreSQL. This helped in avoiding any unauthorized access.
Post-implementation Training and Support :
After the initial consulting and contributions in PostgreSQL setup for NASDAQ X-stream Trading System Platform, Ashnik ramped up the PSE team by providing the required PostgreSQL training.
With first-hand understanding of the PostgreSQL setup and environment, our trainer helped the PSE team learn Postgres with a training program customized for its environment. PSE team, having a good knowledge and understanding of other databases like Oracle and SQL Server was very efficient in picking up the new concepts and handling operational tasks as a part of the training program. Upon completing training, the PSE team was ready to handle their PostgreSQL setup and operations that would host the database for its core system.
- Brought business transformation to PSE’s end users by opting for a new trading platform.
- Increased PSE’s business potential by empowering them to effectively and seamlessly offer and support additional products and services.
- Stable deployment of PostgreSQL, tuned setup and trained staff, helped PSE to focus on its core business instead of worrying about the underlying IT infrastructure of the setup.