Royal Bank of Scotland (RBS)

Client & Industry

The Royal Bank of Scotland (RBS) in Stamford, CT, operated a critical system for allocating Mortgage-Backed Security (MBS) TBA (To-Be-Announced) forward contracts, handling volumes of approximately $20 billion per month. Optimizing these allocations was crucial for profitability.

The Challenge

RBS sought to enhance their existing MBS TBA allocation system ("Tabasco," a C# WinForms application) to improve efficiency and optimize allocation decisions. Key challenges included:

• The need for real-time monitoring of pool deliveries from the Depository Trust & Clearing Corporation (DTCC).
• Implementing complex valuation logic for incoming MBS pools.
• Developing an algorithmic approach ("auto-turn") to identify and allocate the lowest-value pools to fulfill sell orders optimally, maximizing profit.
• Integrating new features into a complex, multithreaded C# WinForms application with a sophisticated backend stack involving C++/CLI wrappers, Tibco Rendezvous messaging, Solaris-based C++ caching servers, and Sybase databases (Open Servers, triggers, stored procedures).
• Enhancing the existing presentation layer (Farpoint Excel spread integration).
• Managing requirements, impact analysis, and task breakdown for new feature development within this intricate environment.

Solution Delivered

Leveraging .NET Technical Leadership and C++ development expertise, Nimbletec executed the following:

• Distill Requirements & Conduct Rigorous Impact Analyses: Meticulously analyzed complex business requirements for enhancing the Mortgage-Backed Security (MBS) To-Be-Announced (TBA) allocation system. Performed comprehensive impact assessments across the existing technology stack, provided accurate time estimates, and enumerated detailed development tasks for the new "AutoTurn" algorithmic trading feature and associated system enhancements within the Microsoft Visual Studio environment.
• Implement Core Auto-Turn Algorithmic Logic: Developed the sophisticated real-time algorithmic allocation logic (AutoTurn) primarily in C# within the .NET Framework. This involved enabling automated monitoring of Depository Trust & Clearing Corporation (DTCC) pool deliveries received via Tibco Rendezvous messaging, implementing dynamic pool valuation algorithms, and executing logic to optimally allocate the lowest-value pools to fulfill sell orders, thereby maximizing profitability.
• Execute Complex Full-Stack Enhancements: Successfully implemented numerous feature enhancements within the existing multithreaded C# Windows Forms (WinForms) application ("Tabasco"). This required augmenting core domain entities and carefully managing data flow and UI thread marshalling across a complex, heterogeneous technology stack involving managed C++/CLI (Common Language Infrastructure) wrappers interfacing with a data access layer subscribed to Tibco messages, Solaris-based C++ caching servers (utilizing Rogue Wave libraries), and a Sybase database backend employing Open Servers, triggers, and stored procedures.
• Enhance Presentation Layer & User Experience: Improved user interface efficiency and usability for traders by enhancing the integration with specialized Farpoint Spread Excel components within the C# WinForms application.
• Develop Supporting Backend Services & Wrappers: Created necessary additional backend services and integration points using C#, C++/CLI managed wrappers, and potentially leveraging existing Rogue Wave C++ library infrastructure to facilitate the new AutoTurn functionalities and data flows.
• Deliver Client-Side AutoTurn Proof of Concept (POC): Provided a complete, functional client-side Proof of Concept (POC) implementation of the AutoTurn feature integrated within the existing application framework. This successfully validated the technical feasibility, effectiveness, and potential business value of the algorithmic allocation approach before full-scale deployment.

Results & Impact

Nimbletec's expert technical leadership and development delivered critical improvements to RBS's high-volume MBS TBA allocation system:

• Implemented Real-Time Algorithmic Allocation (AutoTurn): Successfully deployed the complex AutoTurn feature, enabling automated, intelligent MBS pool allocation driven by real-time DTCC data feeds (via Tibco Rendezvous) and dynamic pool valuations.
• Maximized Profitability on High-Volume Allocations: Directly optimized financial returns on approximately $20 billion per month in MBS TBA allocation volume by systematically identifying and allocating the lowest-value pools to fulfill sell orders through the new algorithm.
• Successfully Integrated Complex Features into Legacy System: Demonstrated expertise in augmenting a sophisticated, multithreaded C# WinForms application and its diverse backend technology stack (C++, C++/CLI, Tibco, Sybase, Solaris) with advanced algorithmic trading capabilities.
• Improved Operational Efficiency & User Workflow: Enhanced the presentation layer using Farpoint Spread components, improving usability and streamlining the workflow for traders and operations personnel interacting with the allocation system.
• Validated AutoTurn Feasibility & Business Value: Confirmed the technical viability, effectiveness, and significant business value proposition of the algorithmic allocation approach through the delivery of a successful client-side Proof of Concept (POC).