Case study
Trade Routing Engine
Bringing order to the chaos through semantic integration, and trade routing
This major European investment bank has complex rules around trade routing and decomposition which determine how risk is hedged, and trades are assigned to their correct ledgers. Their existing solution - a combination of rules, mapping logic and point-to-point connectors scattered across multiple different systems - had become very costly and difficult to maintain.
They engaged Notional to help them define a new strategic solution, and bring order to the chaos migrating their existing rules.
We used a semantic integration strategy to deliver a routing engine with human readable config rules, which saw the typical project delivery time reduce from 5 weeks, to 5 days - a reduction of 80%.
Notional needed to find a way to centralize the routing logic, and provide a self-documenting set of rules, so that users could understand at a glance the end-to-end routing logic that would be applied for each type of trade.
With six independent booking systems involved in trade execution - each with their own API's and booking schemas - we needed to mitigate the risk of a mammoth integration task.
With so many integration points and different API's used to represent the same underlying concept, this was a clear winner for Semantic Integration. We created a platform agnostic, plain english taxonomy to describe the core trade concepts, and then mapped this taxonomy onto the existing API's.
Rather than having to build point-to-point integration solutions, our semantic layer is able to automatically generate trade messages for any of the connected booking platforms. This keeps costs of change low as the booking platforms evolve their API's, or new requirements introduce the need for additional platforms to be added.
We then built out a bespoke DSL to enable the routing rules to be described using the same plain english taxonomy.
Trade decompositions, allocations, hedging rules and mappings. The result was both technical interfaces,
and business logic were now represented in the language of the business - the plain english taxonomy.
Notional also built a robust realtime dashboard, showing requests received, the rules that were applied, the resulting messages delivered, and the responses from downstream booking systems. Integrated into the support tooling for alerting, this has become a central tool in the day to day support cockpit of the banks STP strategy.
We extended the support dashboard to include capabilities to generate regression tests from real trade flows, allowing rapid creation of full regression suites during build & test cycles. This helped keep future releases bug free, without requiring large investment in manual generation of tests.
Future releases will focus on leveraging the DSL to produce auto-generated visual flowcharts, that serve as self-documenting overviews for business users.