Use Case FRTB

Fundamental Review of the Trading Book is a challenging regulation because it requires historical retention of information for back testing of model changes to verify that risk forecasts are consistent with actual data. The standard approach is to warehouse each daily data set separately in a data lake .

Hiperspace solves the problem of historical views with native versioning of elements, accessing As-At is as simple as providing the parameter when opening SubSpace. Tests with tens of thousands of revisions do not significantly impact performance because the latest version is always found first. Using segments and aspects the version history can be confined to only the parts of a model that changes – if only the price or rate changes, only that part needs to be versioned.

Schema evolution allows older versions of financial models to view data exactly as it was when created, while newer versions see any additional aspects added later.

Whole portfolios can be added again to Hiperspace, but only parts that have changed are stored as new revisions.

With unlimited versions, support of arbitrary complex nested objects up to 2Gb in size and transparent partitioning, data volumes do not grow exponentially. It is practical to use a single store for current and historical data.