EAI can be used for different purposes:

• - Data (information) integration: ensuring that information in various systems is consistent. This is also known as EII (Enterprise Information Integration).

• - Process integration: linking business processes between different applications.

• - Supplier independence: extracting the policies or business rules from the applications and implementing them in an EAI system, so that any of the applications used can be changed without said business rules having to be reimplemented.

• - Common Facade: An EAI system can act as the front-end of a cluster of applications, providing a single and consistent access interface to those applications and isolating users from interacting with different applications.

Integration patterns

There are two patterns that EAI systems implement:[2]​.

• - Mediation: Here, EAI systems act as the link of routers between various applications. Where an interesting event occurs in some application (e.g. new information is created, a new transaction is completed, etc.) an EAI system integration module is notified. The module then propagates those changes to the other relevant applications.

• - Federation: in this case, the EAI system acts as an information consolidator between various applications. All access from the external to any of the applications is received by the EAI system and it is configured to expose only the relevant information, connecting to the applications from the external world and carrying out all interactions with the internal applications without intervention of the external agent.

The two patterns are frequently used together. The same EAI system can have several applications in sync (mediation), while serving requests from external agents against those applications (federation).

Access patterns

EAI supports both asynchronous and synchronous access patterns, the first being the usual one in the case of the mediation pattern") and the second in the case of federation").

Patterns life

An integration operation can be "short-lived" (for example, data synchronization between two applications can be maintained in one second) or "long-lived" (for example, in one of the steps it may be necessary for the EAI system to require approval by a human agent of a loan and it may take hours or days to be authorized).

There are two main topologies: hub-and-spoke, and bus. Each of them has its own advantages and disadvantages:.

• - In the hub-and-spoke model, the EAI system acts as the center (the hub), which interacts with the applications, via conversations (or spokes).

• - In the "Bus (Computing)" bus model, the EAI system is the bus (or is implemented as a module resident in an existing message bus or message-oriented middleware).

• - Bus/hub: This is often implemented by extending the functionality of existing middleware products (application servers, message buses) or implemented as a monolithic program") (e.g., without using any middleware), which acts as its own middleware*.*.

• - Application connectivity: The bus/hub connects to applications using a set of adapters (also known as connectors). Those are programs that know how to interact with the specific application. The adapter performs two-way communication, sending requests from the hub to the application, and notifying the hub when an event of interest occurs in the application (a new record is inserted, a transaction is completed, etc.). Adapters can be either specific to the application or to an application set. The adapter can reside in the same process space as the bus/hub or run in a remote location and interact with the hub/bus through industry-standard "Protocol (Computing)" protocols such as message queues, web services, or proprietary protocols.

• - Data formatting and transformation: to prevent each adapter from having to convert data going to or coming from other applications, EAI systems usually employ a common data format, to and from which application formats are converted using transformation services. This is done in two steps: the adapter converts the information from the application format to the common bus format; and then semantic transformations can be applied to this (example: converting zip codes to city names, separating/merging objects from one application into objects from other applications, and so on).

• - Integration modules: an EAI system can participate in concurrent integration operations at any given time, each type of integration is processed by a different integration module. Integration modules subscribe to events of specific types and they receive process notifications when those events occur.

• - Transaction support: when used for process integration, the EAI system provides transactional consistency between applications by executing all operations that involve a single distributed transaction (using the two-phase commit protocol or compensation transactions (operations that undo actions on a given system).[3]​.

In 2003 it was reported that 70% of all EAI projects failed. Most of these failures were not due to technical failures of the software itself, or to the implementation, but rather due to governance problems. EAIIC General Manager Steve Craggs determined the seven main challenges facing companies using EAI systems and explains solutions to these problems.[4]​.

1. • Constant change: The very nature of EAI is dynamic and requires dynamic project managers for its application.

2. • Lack of experience in EAI: EAI requires knowledge of many problems and technical aspects.

3. • Competing standards: Within the field of EAI, the paradox is that EAI standards are not in themselves universal, since each particular provider tries to impose its own.

4. • EAI is a tool paradigm: EAI is not a tool, but a system and must be implemented as such.

5. • Building interfaces is an art: Carrying out the solution engineering process may not be enough. Solutions need to be negotiated with user departments to achieve a common consensus on the final product. The lack of consensus in interface designs tends to lead to excessive effort in mapping the data requirements of various systems.

6. • Lack of detail: Information that at first seems unimportant, over time can become crucial.

7. • Accountability: Since various departments may have contradictory requirements, accountability must be ensured for the final structure of the system.

Other potential problems may encompass the following areas:.

• - Lack of centralized coordination of the work of the EAI.[3]​.

• - New requirements: EAI implementations must be extensible and modular to allow future changes.

• - Protectionism: Applications whose data is integrated frequently belong to different departments which have technical, cultural and political reasons for not wanting to share their information with other departments.

EAI technologies are still in development and there is no consensus on what is the ideal approach or the right set of technologies that an organization should use. The future of EAI will have to result in languages ​​that allow the design of integration solutions at a high level of abstraction, independent of technologies, and in which it is possible, in an automatic way, to make transformations of the solution models to specific technologies.

• - Health Level Seven or HL7.

• - iBOLT").

EAI can be used for different purposes:

• - Data (information) integration: ensuring that information in various systems is consistent. This is also known as EII (Enterprise Information Integration).

• - Process integration: linking business processes between different applications.

• - Supplier independence: extracting the policies or business rules from the applications and implementing them in an EAI system, so that any of the applications used can be changed without said business rules having to be reimplemented.

• - Common Facade: An EAI system can act as the front-end of a cluster of applications, providing a single and consistent access interface to those applications and isolating users from interacting with different applications.

Integration patterns

There are two patterns that EAI systems implement:[2]​.

• - Mediation: Here, EAI systems act as the link of routers between various applications. Where an interesting event occurs in some application (e.g. new information is created, a new transaction is completed, etc.) an EAI system integration module is notified. The module then propagates those changes to the other relevant applications.

• - Federation: in this case, the EAI system acts as an information consolidator between various applications. All access from the external to any of the applications is received by the EAI system and it is configured to expose only the relevant information, connecting to the applications from the external world and carrying out all interactions with the internal applications without intervention of the external agent.

The two patterns are frequently used together. The same EAI system can have several applications in sync (mediation), while serving requests from external agents against those applications (federation).

Access patterns

EAI supports both asynchronous and synchronous access patterns, the first being the usual one in the case of the mediation pattern") and the second in the case of federation").

Patterns life

An integration operation can be "short-lived" (for example, data synchronization between two applications can be maintained in one second) or "long-lived" (for example, in one of the steps it may be necessary for the EAI system to require approval by a human agent of a loan and it may take hours or days to be authorized).

There are two main topologies: hub-and-spoke, and bus. Each of them has its own advantages and disadvantages:.

• - In the hub-and-spoke model, the EAI system acts as the center (the hub), which interacts with the applications, via conversations (or spokes).

• - In the "Bus (Computing)" bus model, the EAI system is the bus (or is implemented as a module resident in an existing message bus or message-oriented middleware).

• - Bus/hub: This is often implemented by extending the functionality of existing middleware products (application servers, message buses) or implemented as a monolithic program") (e.g., without using any middleware), which acts as its own middleware*.*.

• - Application connectivity: The bus/hub connects to applications using a set of adapters (also known as connectors). Those are programs that know how to interact with the specific application. The adapter performs two-way communication, sending requests from the hub to the application, and notifying the hub when an event of interest occurs in the application (a new record is inserted, a transaction is completed, etc.). Adapters can be either specific to the application or to an application set. The adapter can reside in the same process space as the bus/hub or run in a remote location and interact with the hub/bus through industry-standard "Protocol (Computing)" protocols such as message queues, web services, or proprietary protocols.

• - Data formatting and transformation: to prevent each adapter from having to convert data going to or coming from other applications, EAI systems usually employ a common data format, to and from which application formats are converted using transformation services. This is done in two steps: the adapter converts the information from the application format to the common bus format; and then semantic transformations can be applied to this (example: converting zip codes to city names, separating/merging objects from one application into objects from other applications, and so on).

• - Integration modules: an EAI system can participate in concurrent integration operations at any given time, each type of integration is processed by a different integration module. Integration modules subscribe to events of specific types and they receive process notifications when those events occur.

• - Transaction support: when used for process integration, the EAI system provides transactional consistency between applications by executing all operations that involve a single distributed transaction (using the two-phase commit protocol or compensation transactions (operations that undo actions on a given system).[3]​.

In 2003 it was reported that 70% of all EAI projects failed. Most of these failures were not due to technical failures of the software itself, or to the implementation, but rather due to governance problems. EAIIC General Manager Steve Craggs determined the seven main challenges facing companies using EAI systems and explains solutions to these problems.[4]​.

1. • Constant change: The very nature of EAI is dynamic and requires dynamic project managers for its application.

2. • Lack of experience in EAI: EAI requires knowledge of many problems and technical aspects.

3. • Competing standards: Within the field of EAI, the paradox is that EAI standards are not in themselves universal, since each particular provider tries to impose its own.

4. • EAI is a tool paradigm: EAI is not a tool, but a system and must be implemented as such.

5. • Building interfaces is an art: Carrying out the solution engineering process may not be enough. Solutions need to be negotiated with user departments to achieve a common consensus on the final product. The lack of consensus in interface designs tends to lead to excessive effort in mapping the data requirements of various systems.

6. • Lack of detail: Information that at first seems unimportant, over time can become crucial.

7. • Accountability: Since various departments may have contradictory requirements, accountability must be ensured for the final structure of the system.

Other potential problems may encompass the following areas:.

• - Lack of centralized coordination of the work of the EAI.[3]​.

• - New requirements: EAI implementations must be extensible and modular to allow future changes.

• - Protectionism: Applications whose data is integrated frequently belong to different departments which have technical, cultural and political reasons for not wanting to share their information with other departments.

EAI technologies are still in development and there is no consensus on what is the ideal approach or the right set of technologies that an organization should use. The future of EAI will have to result in languages ​​that allow the design of integration solutions at a high level of abstraction, independent of technologies, and in which it is possible, in an automatic way, to make transformations of the solution models to specific technologies.

• - Health Level Seven or HL7.

• - iBOLT").