Intro to Integration Architecture Patterns

An easy way to understand the difference between various integration architecture patterns is to imagine that your data is cheese.

Have you ever wondered what if your business data turns into cheese? Yes - actual cheese. No - you have not probably done that. This blog compares common integration architecture patterns and approaches a situation where your data is cheese.

  1. Once in a while, you just need cheese. You stand up and get it from a refrigerator. This is a point-to-point integration, and it is efficient as long you are alone eating the cheese.
  2. The whole family wants cheese as a part of the lunch. Family members do go one by one to get it from a refrigerator and return it. One member gets the cheese and delivers it to each one. This is called an integration via a centralized integration platform. That member who provides the cheese can monitor and control the cheese consumption of the whole family.
  3. The extended family gathers together for dinner. The cheese is on the table, and everyone who wants cheese, can grab and take it. It is easy to control and monitor the consumption of cheese. This is integration via API as the cheese is easily available for the family members that want it.
  4. If one family member constantly reserves the cheese platter, thus disabling the access from other family members, it is a Denial-of-Service attack on your chee..API. If two family members agree that they will hoard all the cheese by themselves, it is a Distributed Denial-of-Service attack on your service.
  5. There are all the varieties of cheeses available on a cheese platter. Dinner members choose what they want — this is a Service-oriented-architecture as everything that you have is available. Dinner members are your systems or processes that require that chee.. data via APIs.
  6. You add cheese snacks on the platter. They include everything and are easy just to take and fit perfectly in your mouth. They don't even require wine - they're just perfect as they are. These are microservices - fully independent and just the right granularity. The frustrating thing is that they require more work than the other approaches. They're currently very popular, but do have also their own pitfalls.
  7. Your family starts to produce cheese. The delivery of your delicious cheese is made by a delivery company to shops. This is a B2B integration, and your integration platform is the company making the deliveries the ones who use your information or resell it as a part of their offering.
  8. Your cheese factory has its own branded display cases inside the stores. Cheeses are neatly described and sold in industry-standard packages. Buyers can even have test tastes of the cheese. This is an OpenAPI as the users of information can easily access well-defined data, and there are test messages available. The display case is your API portal. You are also monetizing your data.
  9. Third-party lunch service is offering your cheese as a part of their meals. Meals contain a set of food elements produced by multiple parties, and delivery is included via yet another party. This is API-economy: the third party uses your's and others data and services together to create a new service.
  10. You start to use automated factories, quality assurance warehouses and delivery systems for your cheese. This is the CI\CD pipeline of your business.