Canonical Data Model - the solution to allow re-use in complex integration landscapes

Voordat ik het Canonical Data Model uitleg, zal ik een korte introductie geven die kan helpen begrijpen waarom organisaties deze oplossing nodig hebben. Door de digitale transformatieprogramma’s waar bedrijven aan beginnen, zien we dat integraties een steeds belangrijker onderdeel worden van IT-budgetten. Volgens Gartner zal 50% van alle IT-budgetten in 2020 worden besteed aan het integreren van systemen en processen. De trend om de beste oplossingen te gebruiken door de opkomst van low-code applicatie-ontwikkeling betekent dat het gemakkelijker wordt om specifieke processen binnen een organisatie op te lossen door de ontwikkeling van afzonderlijke applicaties.

This development results in an increasingly fragmented IT landscape, which ultimately leads to severe integration issues: what information has to be available where and when? Which data is leading? Do the applications speak the same “development language”?

Point to point connections

With point to point connections, applications are directly linked on a one on one basis. Data is exchanged between these applications. Logically, the amount of point to point connections with multiple applications in your landscape will increase exponentially, as you can see in the graph on the left below.

Whereas with 3 systems the number of point to point connections is still three, with 10 systems this numer has already increased to 45 connections. A recipe for spaghetti!

Not every application needs to be integrated with every single other application, but the graph clearly shows the impact of adding a 12th application. The negative impact on time, money and quality by adding or replacing applications in the point to point scenario is nearly disastrous. Besides that, the risk of mistakes and instable network traffic are high.

Structure your complexity: scale expansion with a Canonical Data Model

So how do you scale up your integrations landscape in a proper way? By using a Canonical Data Model. A Canonical Data Model is often referred to as the common data model: one language that connects all applications. Thus enforcing standardization of data definitions. Therefore, this Canonical Data Model wil hold the single integration model for your company – one truth.

Canonical Data Model: key advantages

• Easy overview: All integrations are implemeted in a single model (the Canonical Data Model). There is one location where you can see all involved systems and integrations.
• One single language: All applications are translated back to a single model: the Canonical Data Model. Interpretation of data is managed on a central level.
• Scalable: Adding or replacing applications means translation only to the Canonical Data Model, in stead of to all other applications and systems. The Canonical Data Model will manage this automatically on a central level.
• Speed of development: Connecting applications to the Canonical Data Model is a single operation. Connecting aplications point to point are multiple operations, which costs more effort and time.

Key considerations when setting up a Canonical Data Model

There are a few considerations to take into account when designing your Canonical Data Model:

• Initial investment: The setup of a Canonical Data Model will take more time and effort at start-up time. But after three applications, it is becoming easier and the relative time and effort per integration will drop dramatically.
• Implementing and maintaining a Canonical Data Model is not a rush job – doing the job first time right results in direct benefits.
• Considering future evolutions in your integration landscape so you can already take that into account when designing your future proof Canonical Data Model.
• A Canonical Data Model remains custom for every organisation despite the fact that there are many standard connectors for applications and systems.

Simplify replacing Legacy systems!

When replacing legacy systems or adding new systems or applications, you only have to connect these to the central Canonical Data Model. If you continue to focus on the goal to have a scalable integration landscape, you will find out that the Canonical Data Model provides that overview and scalability.

Thanks for reading this article about the Canonical Data Model. Don’t hesitate to contact us in case you have any questions or want to know more!

By Alexander Willemsen, CTO @ eMagiz