A Tervela Case Study: Ensuring Data Consistency Across Global Trading Systems
Trading Applications Need Consistent, Continually-Updated Information
One of the most critical requirements for a high-speed trading system with a distributed, co-located architecture is that each trading application needs to know what the firm’s overall position is in each symbol traded, in real time. This requirement is driven by two essential functional areas:
- Risk Management: Each firm must know what its net position is on a per-symbol and a total cash basis continuously in real-time. When that isn’t the case, either due to slowness or data loss, the firm may have inconsistent positions resulting in the firm inadvertently violating its risk limits. In extreme cases, this leads to a violation of a firm’s maximum leverage, and results in disciplinary actions by a firm’s prime broker such as the requirement for additional capital, forced liquidation of positions, or suspension of trading.
- Compliance: Each firm must have a record of all order activity including orders, modifications, cancels and fills. If a firm can’t produce data proving this compliance, it can be censured, fined, or worse.
Tervela’s customer, in the midst of substantial growth in its high-frequency trading operations, was struggling to maintain an acceptable degree of global data integrity with their existing messaging infrastructure.
Dropped Packets Led to Inconsistent Information
Tervela’s customer needed to revamp its position management system to respond to the mismatch that existed between the business risk of the regulatory environment and the systems in place to manage those risks. Specifically, they needed a way to consistently update multiple databases and trading engines with every order, modification, cancel, and fill from each of their globally distributed trading systems. However, they were often experiencing inconsistently recorded positions – especially across geographically dispersed instances. This was leaving the company exposed to fines and the other costs and risks already described.
These inconsistencies were attributed to the fact that their legacy systems employed multicasting protocols to transport each update to all co-located trading systems. Geographically dispersed systems will inevitably experience packet loss when using multicast – which is exacerbated during periods of intense trading or messaging. In these environments, high-volume firms in particular can have a disproportionally large problem because dropped packets result in lost information, which in this case translated to the inconsistent positions they were seeing across regions.
Achieving Consistency and Performance with Tervela’s Data Fabric
Tervela delivered the missing piece they needed: a high-performance, no-loss data fabric, that excelled in both LAN and WAN-connected topologies, and gave the firm the ability to persist and replay a month’s data by symbol (topic), source, and/or time period.
Benefits of the Tervela-enabled Solution
With Tervela, position management systems are always consistent with each other and never have gaps in trade data due to loss.
- Avoid Fines: SEC fines and other severe penalties for mismarked orders and inconsistent positions are eliminated, and the risk associated with them becomes adequately managed.
- Reduced Operational and Compliance Costs: Tervela customers avoid the need to allocate significant human capital to perform the forensic work needed to explain gaps in data and mismarked positions. Operations are no longer interrupted by the compliance process.
- P&L Performance: The ability to distribute data quickly and without loss between critical applications means that all the systems are more coordinated, and the gap between trading strategy and execution is reduced
- Time-to-Market: The Tervela data fabric provides a ready-made infrastructure for mission- critical applications, dramatically reducing the time to develop, test, roll out, and manage them. Reducing this cycle time gives Tervela customers a competitive edge in the market.
Selection Criteria
This customer selected Tervela because of the following differentiated advantages:
- No-loss architecture: A system that drops messages does not eliminate the risk of fines and sanctions for mismarked orders and inconsistent positions. Tervela’s message delivery proved to be the most advanced, maintaining the highest degree of data integrity.
- Highest Availability/Reliability: If the position management system is down, the firm has to stop trading, costing hundreds of thousands of dollars in lost profits an hour. After extensive due diligence, this customer determined that Tervela was the most resilient product.
- Vendor expertise and responsiveness: This customer rarely buys technologies from 3rd parties, but when it does it looks for a vendor that can be an extension of its own team. Tervela fit that requirement, delivering the most responsive support in the industry.
- Enabling functionalities: Tervela’s integrated persistence and seamless replay features solved data ingestion and loss problems associated with surges that typically affect downstream consumers, such as high throughput trading engines or records databases.
Conclusion
As regulatory pressure and data volumes continue to increase, firms will be forced to make significant improvements in the infrastructure that support the middle and back office. Increasingly, their infrastructure must combine the following five attributes: 1) real time, 2) no-loss, 3) high- throughput, 4) global, and 5) high availability.
Firm after firm is realizing that, given the combination of time-to-market pressures that the regulatory environment imposes, the scarcity of highly trained and affordable human capital, and the complexity of building no-loss, high-availability global systems, it no longer makes sense to build an enterprise data fabric in-house. Tervela enables trading firms to re-platform the no-loss data movement functionality of critical systems like real-time risk management more cost- effectively, with higher reliability than has been possible in the past.
< Back