What is an Integration and types?

‘Integration’ is the process of combining two or more things to create a whole. For businesses, the term integration often refers to software or system integration, which means bringing together multiple business systems to operate as a collaborative unit.

At the simplest level, integration is the process of combining two or more things to create a whole. For businesses, the term integration often refers to software or system integration, which means bringing together multiple business systems to operate as a collaborative unit.

Integration allows information to be shared between the connected systems. These integration solutions can come in many forms, whether it be requesting information from a website, internal employee systems sending and receiving information, or connecting customer data from a point of sale system to CRM to automate recommendations.

Regardless of industry, software integration can be applied to several use cases for nearly every aspect or department of the business, so long as there are multiple systems, applications, devices, or datasets that need to be connected.

Types of integrations:

Point-to-Point Integration (1–1 Mapping): This integration is ideal for one business function at a time, typically connecting just one system to another. Integrating more systems together with this method results in more one-to-one connections and quickly becomes unmanageable.

P2P pattern.

Vertical Integration — This integration differs from other types of integration in that one sub-system is linked to another sub-system based on how closely they are related with respect to the functions that they each perform. This creates a silo-like structure with the bottom sub-system being the most basic and the remaining sub-systems getting progressively more complex as the structure is ascended. This integration is inflexible and unmanageable as adding more functionalities results in more silos created in the system.

vertical integration

Star Integration (Many to Many Mapping)— This type of integration is a more complex version of the point-to-point integration. A large number of systems may be connected together to form many-to-many relationships with crisscrossing “spaghetti-like” connections. The increased functionality over point-to-point integrations comes at the cost of complexity, and star integrations can quickly become unwieldy to manage.

Star integration

Horizontal (ESB) Integration — This type of integration is achieved by using one specialized sub-system as a common interface layer (or main system) that connects to all of the other sub-systems. In this integration, the number of connections is reduced because systems are not directly connected to one another. Instead, they are indirectly connected by way of a common main system which serves as a conduit, broker, and hub. The main advantage of this method is the lower number of connections needed to maintain functionality, thereby reducing the amount of time, effort, and resources invested in building the system. This intermediary layer/subsystem used in this form of systems integration is commonly known as an Enterprise Service Bus (ESB).

Horizontal (ESB) Integration

Types of integration patterns:
Migration: Data migration is moving a data set from one system to another. Think of this as a data transfer for a certain point in time, meaning that this exact data set will only be moved once. This occurs commonly when businesses switch from one system to another and the data in the old system needs to moved over.
Broadcast: A broadcast pattern is often referred to as one-to-many communication when one system sends data across many systems. Broadcast integration involves real-time updates or updates that occur on a regular cadence. This pattern is used to keep information across systems as up-to-date as possible. Aggregation: An aggregation integration pattern can be thought of as many-to-one communication between systems. It keeps data from multiple systems in sync and up-to-date within a single system, allowing businesses to merge data and have a singular view of the data from multiple systems on demand.Bidirectional sync: Bidirectional sync is when two or more integrated systems communicate back and forth and behave as one system with multiple data sets. This is used for systems that serve their own business purpose but can accomplish even more when connected to related real-time data sets. Correlation: A correlation integration pattern is similar to bi-direction sync in that it is placing real-time data in multiple systems, however, it does not replicate data only associates it to similar data. If data exists in two systems, this type of pattern will indicate that, but if the record only exists in one system it will not be created.



Next > Introduction to MuleSoft.