Apache Tuscany > Home > Java SCA Subproject > Java SCA Menu > Java SCA Architecture Overview |
This page is work in progress. Thanks for your contribution
At the moment we are pulling different pieces of information in one place and then we'll decide on how to organize the information.
The SCA Java runtime is composed of a kernel and extensions. The kernel is essentially a multi-VM wiring engine that connects components together using the principles of Dependency Injection, or Inversion of Control.
The kernel is designed to be simple and limited in its capabilities: it wires functional units together. Capabilities such as service discovery, reliability, support for transport protocols, etc. are provided through extensions. The kernel extension model is based on the SCA 1.0 Java programming and assembly model. In otherwords, the kernel uses SCA to bootstrap itself. The mechanics of how this is done are described in detail in Kernel Architecture Guide.
The kernel is divided into a number of modules: api, host_api, spi and core. The api module defines Tuscany proprietary programming model annotations and APIs. The host_api module defines APIs for interacting with the kernel. The spi module defines extension points and abstract helper classes. The core module contains the kernel implementation.
Extensions enhance the kernel with specific sets of functionality. Extension types are not fixed and the kernel is designed to be as flexible as possible by providing an open-ended extension model. From the kernel perspective, extensions are just a set of SCA components. However, there are a number of known extension types defined by the spi module, including:
There are also a number of add-on extensions that do not correspond to a well-know type:
More detail can be found in the Extensions Guide.
The kernel is designed to be embedded in, or provisioned to, a number of different host environments. For example, the kernel may be provisioned to an OSGi container, a standalone runtime, a servlet engine, or JEE application server. There is even a Maven runtime that embeds kernel for integration testing. Runtime capabilities may vary based on the host environment.
Not sure if this belongs to this page or not
Data provides transparent data mapping between different types of data that can be exchanged between caller of a service and the service itself given the format of the data that will be supplied by the client. Data Mediator will do this transformation given the following information:
You can find more detailed information about databinding at this link