Rosetta Engine

Rosetta Engine is a run-time execution engine that can run the functional logic contained in a model on physical data. Once this functional logic has been modelled, Rosetta Engine transforms it into an executable data pipeline using the source code that is automatically generated from that model.

The Rosetta Engine is provided with a user interface in the Rosetta web application and also as an API. The API takes model objects as inputs, executes the data pipelines and returns the output. This API can be deployed either on cloud or on premise.

In combination with other components of the Rosetta Platform (e.g. Rosetta Design, Ingest and Monitor), Rosetta Engine supports the regulatory reporting capability of the platform. When used for regulatory reporting purposes, the API can execute regulatory rules to create the reports required by regulators from their regulated entities.

Main characteristics

  • Generic: Rosetta Engine can transform any data processing model into a physically executable data pipeline. The engine is made available under a comprehensive set of programmatic languages.
  • Live: Rosetta Engine can execute a data pipeline in real-time as user define their data processing model. This enables a test-driven approach where a model is developed together with a set of test scenarios.
  • Scalable: Rosetta Engine scales based on the number of data objects that need to be processed.

Digital Regulatory Reporting

The Rosetta Platform provides a version of the Rosetta Engine that has been customised for regulatory reporting purposes and supports an industry process called Digital Regulatory Reporting (DRR). In DRR, industry participants use the Rosetta Platform to digitise the framework for reporting and collecting regulatory data, therefore mutualising the interpretation of reporting rules and their technical implementation.

The DRR approach is based on 3 principles to implement regulations:

  • model-driven
  • functional
  • test-driven

Under this approach, regulatory reports are defined as data pipelines. Each pipeline consists of a set of regulatory rules defined as logical instructions (i.e. functions) applied to model objects. Rosetta Engine provides the pipeline’s execution engine that produces the regulatory reports to be consumed by regulated entities and regulators. Those reports are shown in the Reports Feature that is accesible from the bottom banner.


The next section describes how users can practically apply the 3 principles of DRR to build regulatory reports.

Step-by-step approach for regulatory reporting in Rosetta

Step 1: Develop a model

The initial step consists in building a regulatory-agnostic model for the business process or activity being regulated. This model supports the definition of the objects that will be used as inputs to create the report.

In the below example based on the Demonstration Model, the Definition View in Rosetta Design is showing the components of the model filtered by Type.


Step 2: Represent regulations as functional expressions

The process to define a regulatory report has 3 parts:

  1. Define and organise the set of regulatory documents that will be referenced by the rules
  2. Define rules as functional expressions that reference specific provisions in those documents
  3. Define a report by collecting and chaining those rules together

The Rosetta DSL provides a simple and intuitive syntax to represent those 3 components. Complete information about that syntax can be found in the Reporting Components section of the Rosetta DSL documentation.

1. Start from the regulatory text

First, in order to define a regulatory report in the model, the set of regulatory documents that will be referenced by the rules need to be accessed, with extraction of key information:

  • The entity that has issued the regulation, represented as a body in the regulatory model
  • The document defining the regulation itself, represented as a corpus in the regulatory model

Both are referenced throughout the report definition.

2. Define rules as functional expressions with references to regulatory provisions

Each reportable field contained in a corpus can be defined by its references to specific provisions and by the reporting rules, represented as functional expressions, that will generate the corresponding reportable value from a model object.

The definition of a reporting rule in the regulatory model contains the following elements (with corresponding syntax):

  1. Field name and description: reporting rule <Name>
  2. Regulatory reference indicating body, corpus and location in the regulatory text, and the provision text: [regulatoryReference <Body> <Corpus> <Segment1> <Segment2> <…> provision <”ProvisionText”>]
  3. The functional expression used to generate the reportable value: extract <Expression>, filter when <FunctionalExpression> etc.
  4. The field name used when visualising the report: as <FieldName>

The result of this can be visualised in the Regulation View of Rosetta Design, shown below for an example based on the Demonstration Model for EU emission standards for vehicles. In the model itself, each rule ties the functional business logic (as shown on the right) to specific document references, including best practice and guidance (as shown on the left).

3. Define a regulatory report by assembling a set of rules

A regulatory report is defined as a data pipeline consisting of a collection of rules that together produce the reportable values required by the regulation. The ordering and naming of those rules aim to reproduce how reportable fields are defined in the relevant corpus.

The definition of a regulatory report in the regulatory model contains the following elements (with corresponding syntax):

  1. Name under which to report the fields: report <Authority> <Corpus1> <Corpus2> <…>
  2. When to report: in <TimingRule>
  3. Whether to report: when <EligibilityRule1> and <EligibilityRule2> and <…>
  4. What to report: with fields <FieldRule1> <FieldRule2> <…>

An example based on the Demonstration Model illustrates the creation of a report called EmissionPerformanceStandardsEU with reportable fields such as Vehicle Registration Id, First Registration Date, Vehicle Classification Type etc.


Step 3: Test regulatory reporting on scenarios

Once the user defines a regulatory report in the model, they can generate reports in real-time and visualise the results in the Reports Feature of the bottom pannel. Each report is executed on pre-defined set of test scenarios that represent input data into the reporting process and are assembled into a test pack.

The following example based on the Demonstration Model shows the output of the EmissionPerformanceStandardsEU report based on one test pack that contains two test scenarios. The reported values, displayed as columns, each correspond to a reportable field defined in the What to report section of the report model.


The test scenarios are a key component of the test-driven approach enabled by Rosetta. Based on these test scenarios, users can validate whether the regulatory report is generated with the expected reportable values. Rosetta allows users to develop those test scenarios by synthesizing input data from the model, on which the regulatory logic can be tested in real-time.

Each test scenario with its input values can be visualised by clicking on the test scenario name (as a hyperlink). The below example displays the values available in one test scenario associated with the test pack used for the above regulation.