The risk calculation process starts with ingesting data from different sources, extracting, transforming, and loading (ETL) it into a distributed cloud for risk calculation. Some data and events are processed in real time, using techniques such as complex event processing and streaming analytics to synthesize external events into business-relevant market events used to trigger new risk calculation processes. The calculations evaluate the risk profile of assets and portfolios, taking into consideration various market scenarios using risk models and algorithms, typically as largely parallel computations. This can happen in a single step, or as a complex sequence of risk calculations that are coordinated via automated business processes. Risk calculations can be scheduled or performed in real time.
The input data is also moved to historical store while the risk calculation results are stored in the result repository. Business rules and ML algorithms are applied to the results, both to derive insights and to identify and automate reactions to certain conditions, improving the responsiveness of the risk management process. The interaction between the compute environment, business processes, and external data feeds is event driven, which ensures the modularity and composability of the process.
Download Red Hat market and credit risk calculation framework architecture Whitepaper
