An introduction to the ontology-derived Enterprise Data Model for Finance and Business stakeholders.

The kickoff and first lesson of the FIBO Data Model Education course introduce the ontology-derived Enterprise Data Model. The first class shows how the Configurable Ontology to Data Model Transformation (CODT) derived the data model from the industry-standard. It explains the differences between the open-source Core and the full Commercial version.

You can read the presentation or download the PowerPoint here.

Transcript of the presentation

Hello and welcome to the FIBO Data Model training. My name is Jurgen, and I’ll be teaching the course and also providing answers to your questions and guidance.
The first session, Semantics for Manager, is basically a non-technical introduction to the ontology-derived data model. The education course has nine classes, usually lasting one to two hours, sometimes longer if it has a lot of follow-along and Q&A.

In addition, there are FIB-DM articles and tools like MS Visio stencils and an Excel workbook. Here, the first lesson is for nontechnical audiences, and you will see that it’s a very high-level lesson that also addresses the open source. But we can go through this very fast.
And you should keep it as a resource. With all the PowerPoints you can copy from it, you can use them for your internal presentations, design, and requirement documents.

Your financial institution already embraces model-driven development, industry standards, and reference models.
And here, “Semantics for …”: that’s a theme throughout the course.
What I mean is that the FIBO is a semantic model, and we leverage its semantics for the wider organization, in particular, when designing databases for core operational systems.
This lesson is for finance stakeholders, subject matter experts, and those with a working knowledge of ER and ontology diagrams, and, of course, for data and application architects, and even for ontologists who can leverage the model and even reverse-engineer data. models into RDF.

About myself, briefly. I worked in the Finance industry for 20 years, and in particular, I was an IBM Software Group consultant for the Banking and Financial Markets Data Warehouse. And in that capacity, I trained Financial Institutes on the model at some 45 banks in North America, Europe, and Asia. After IBM, I advised Citibank and Deutsche in implementing the BFMDW.
I was also a speaker at FIBO conferences from the beginning.
My company, Jayzed Data Models, was incorporated in 1999, and Jayzed has the FIB-DM copyrights and is a designated assignee of the CODT patent.

There is a chasm between semantic and conventional data management. The EDM Council specified the FIBO in Ontology Web Language (OWL).
But OWL needs highly specialized ontologists.
The FIBO model is very comprehensive. It has detailed coverage of business entities, loans, securities, derivatives, indicators, and everything. However, many banks and investment managers don’t have the expertise in-house yet.
Large financial institutions started the implementation of RDF stores. However, IT departments must still support and design conventional databases.
FIB-DM is the bridge across that chasm.

The ontology transformed into a data model leverages the design for relational databases. So, we have the semantic world with the Triple Stores.
And we have 90% of the Financial Institution still running on relational databases.
And now, new is here, we have the FIBO in RDF/OWL, and we have data models to design the RDBMS.
And then, with CODT, the Configurable Ontology to Data model Transformation, the industry standard becomes available as a data model – and that is FIB-DM.
Some predictions: RDF stores dominate data systems for knowledge management. However, RDBMS remains the standard for core transactional BI systems.
Hence, Financial Institutions still need relational models and data modelers.
Now, however, the Ontology Web Language (OWL), I believe, will replace the entity-relationship model as the notation of choice for industry domains and enterprise models.

The reference architecture is SEIA, Semantic Enterprise Information Architecture.
Now, if we look at this, we have the use of design artifacts, be it for business, design, or development.
The type is a model, conceptual, logical, physical, was it an implementation?
And then the level, whether it’s enterprise, departmental, or for a specific project.
What we have had for years was a Data Model for RDBMS, and also XSD, BPMN, and Object Models.
Now, what’s new is the FIBO, RDF/OWL, and RDF stores.
And with FIB-DM, we have the FIBO available as a data model, and we can generate UML models and XSDs.
And finally, there are Concept Maps. That’s a higher level of abstraction for business users who will find concept maps easier to understand and comprehend than data models or ontology graphs.

With that said, semantic model-driven development means, on a conceptual level, FIBO is the Finance Domain Ontology, and FIB-DM is a conceptual data model, the sibling of the FIBO.
It can be used as a reference model for the enterprise and for projects.
And from FIB-DM, we derive Logical Models for data, messages, processes, and objects, and then we can physicalize the models.
Now, what’s important here is that the ontology is at the apex of MDD and of enterprise architecture.
Midsize Financial Institutions without semantic technology centers yet, adopt FIB-DM as a strategic enterprise model.
Large institutions use CODT, the transformation technology, to transform the in-house ontologies into data models for downstream implementation.
The Financial Industry Business Data Model means the FIBO in PowerDesigner or other data modeling tools.
So the model comprises, in its latest release, over 3,000 entities and has complete definitions, annotations, axioms, and business rules.
Data architects leverage the full content of the industry standard.
We have a common language and design patterns for semantic and relational data.

Some of the principles that underlie FIB-DM and considerations.
The model must be practical, and it must be complete.
We do not want to miss any information from the ontology. And there it is, really comprehensive. Everything in the FIBO is available as entity properties in the data model. The model maps back to its source, the ontology. That is also for every entity; we have a URI that points to the FIBO class.

The domain ontology generates a perfect conceptual data model.
On the left-hand side, we have an ontology graph showing a bank account identifier that identifies a deposit account, and the deposit account is provided by a bank, a financial institution.
And we see here that the entity-relationship diagram is the best representation of the bank, the account, and the identifier. So there are no missing or superfluous entities or relationships in the design.

How did we get from the FIBO to FIB-DM? How does it work?
The Configurable Ontology to Data model Transformation (CODT) is basic ETL (Extract, Transform, Load).
We extract metadata from the source ontology, transform it into Conceptual Data Model metadata, and load it into the data modeling tool, PowerDesigner.
And the extraction process runs SPARQL queries on the ontology platform to retrieve the metadata. PowerDesigner imports MS Excel workbooks, and the transformation in between is a two-step process using the patented Metadata Sets.

The CODT Metadata Sets: There’s an extraction process that populates Ontology Metadata Sets for classes, object-, data properties, and annotations.
Then step one transforms Ontology Metadata and populates a generic Entity Relationship Metadata Set.
And then in step two, the generic E/R gets transformed into tool-specific meter data sets, and in this case, these are PowerDesigner object properties and extended attributes.
Now, we serialize to MS Excel and load it directly into the data modeling tool.
The CODT Utility Patent is quite groundbreaking and broad. It means that Jayzed Data Models has protected its intellectual property, and with that, it protects open-source Core users and licenses of the Full Model.
It enables Institutions to transform very large ontologies and make the transformation process available for licensing.
With that, we can transform other domain ontologies like Gene, Oil, and Auto in the future.

FIB-DM has an open-source Core version licensed under the GPL 3.0, a copyleft license, meaning that any transformation or derivative work must also be licensed to the public.
Here are the open-source and commercial models side by side.
The open-source has over 1,000 entities as of Q4 2018, and FIB-DM, in its latest version as of Q4 last year, has 3,173 entities.
FIB-DM Core is a self-contained, standalone data model.
The free open-source includes the generic modules like SKOS, Specification Metadata, and OMG Languages, Countries, and Currencies.
And the domain core, these are the core FIBO modules: Foundation, Finance, Business & Commerce, and Business Entities.
The commercial license: Number one, it includes the OMG Commons Ontology Library, which has replaced many entities in the domain, particularly within the FIBO Foundation.
We get additional modules for Loans, Processes, Securities, Corporate Actions, Derivatives, and Indicators.
FIB-DM Extended, the full version. Financial Institutions have access to the complete 3000-entity data model, which includes quarterly updates.
Every 3 months, the EDM Council releases a new version of the FIBO, and usually four weeks later, there is a new version of the FIBO Data Model
Also, the license grants access to data diagrams, PowerPoints, Excel, and Visio resources. That includes permission to lift off and rebrand the resources for internal use.
The optional 12-month maintenance agreement also includes training model and resource updates.
It’s priced by tier based on the asset size of the financial institution.
The goal is real transparency for your FIB-DM evaluation. The Fremium approach means you can download an open-source version and then decide whether you want to upgrade to premium.
You can review the education resources, which include presentations and articles.
You can examine the full model report, which lists all 3,200 entities with their definitions.
You can then schedule an online call with me to show you a demo and address your questions.
Thank you.