To enable full-scale modeling of integration, Metada Integration Modeler provides features covering the complete integration project lifecycle. The lifecycle includes not only support for design of integration services, but also their validation, autotesting, versioning, synchronization, and deployment.

• Web-based environment

Any number of integration analysts may take part on development and maintenance of one central integration model or even multiple integration models at the same time. User authentication is either via enterprise LDAP or global OpenID. There is no need to install and maintain any special software on user machines - web-browser is sufficient.

• Integration services modeling

Service interfaces are modeled as composable and reusable types. Each type may define tree-structure of data and may be composed of other types. This allows for flexibility and ease of interface definition, while enabling effective reuse. Both XML and non-XML data formats are supported via generic transformations from/into XML. 

Integration services decide what backend services will be called using content-based message router. When constructing a message to be sent, a message transformation may be applied. Transformations include mapping from one tree-structure of data to another tree-structure of data. Different mapping functions and value-mapping tables may be applied during a transformation. Mapping collections of data is supported.

• Model validation

Conformance of integration models to specific constraints may be regularly checked. Validations ensure production quality and consistence of models. Validations check compliance to conventions or structural constraints that need to be enforced.

• Model versioning and synchronization

In enterprise application integration solution it is necessary to support effective long-term development and maintenance. In production model (the one that defines the current state of the integration), only necessary fixes are usually being realized. Most of the development and changes are performed in development models, which may be migrated to production only after thorough testing. For this reason Metada Integration Modeler supports multiple versions of models that may be synchronized among one-another.

• Services autotesting

Each service accepts a message, does some deciding what other services to call, and does some data transformations and message formatting. All this is modeled in the integration model and needs to be repetitively tested. It is necessary that the whole integration model be regularly retested after any sort of changes. In practice it is very expensive to retest systems manually. Metada Integration Modeler provides autotesting feature that allows several automated tests to be specified for each service (usually to test various use cases). The autotest suite may be then executed at any time to check how changes in model have affected only the right integration functionalities. Autotests enable a strong grip on the integration solution.

• Code/configuration generation

Metada Integration Modeler currently features full code/configuration generation for BEA AquaLogic Service Bus (XQueries and configurations) and for BEA Tuxedo (C++ code). Code generation for other integration/orchestration platforms is possible.