Panel |
---|
borderColor | #ECF4D1 |
---|
bgColor | #ffffff |
---|
titleBGColor | #ECF4D1 |
---|
title | Apache Tuscany SCA Java Architecture Guide |
---|
borderStyle | solid |
---|
| |
Note |
---|
| title:Notification |
---|
| title:Notification |
---|
| Center |
---|
This page is under construction- You are welcome to help and complete it |
| OverviewThe SCA Java runtime is composed of core and extensions. The core is essentially a multi-VM wiring engine that connects components together using the principles of Dependency Injection, or Inversion of Control. Anchor |
---|
| Core Definition |
---|
| Core Definition |
---|
| CoreThe Core is designed to be simple and limited in its capabilities. It wires functional units together and provides SPIs that extensions can interact with. Capabilities such as service discovery, reliability, support for transport protocols, etc. are provided through extensions. Anchor |
---|
| Extension Definition |
---|
| Extension Definition |
---|
| ExtensionExtensions enhance SCA runtime functionality. Extension types are not fixed and core is designed to be as flexible as possible by providing an open-ended extension model. However, there are a number of known extension types defined including: - Component implementation types, e.g. Spring, Groovy, and JavaScript
- Binding types, e.g. Axis, CXF, AMQP, ActiveMQ, JXTA
- DataBinding types, e.g. JAXB, SDO, XmlBeans
- Interface Binding types, e.g. WSDL, Java
Details of how to implement an extension can be can be found in the Extensions Guide. Anchor |
---|
| Runtime Definition |
---|
| Runtime Definition |
---|
| RuntimeThe core is designed to be embedded in, or provisioned to, a number of different host environments. For example, the core may be provisioned to an OSGi container, a standalone runtime, a servlet engine, or J2EE application server. Runtime capabilities may vary based on the host environment. High level overview of Java SCA runtimeThe diagram shown below is a high level view of the SCA runtime which consists of the following key modules/packages: - SCA Spec API: The APIs defined by the SCA Java Client and Implementation Spec
- API: Tuscany APIs which extend the SCA Spec APIs
- Core: The runtime implementation and the SPIs to extend it
- Extensions:
- Container implementation - For extending language support, for example BPEL, Python, C++, Ruby,..
- Binding - for extending protocol support, for example Axis2, CXF,..
- Interface Binding - for extending types of service definition, for example WSDL, Java, ...
- Databinding - for extending data support, for example SDO, JAXB, ...
- Host Platforms: The environments that host the Tuscany runtime.
Image Added Anchor |
---|
| Internals High Level View |
---|
| Internals High Level View |
---|
| Internals High Level View[Note: Do we want to link to Kernel-Structure]? BootsrapBootstrap process is controlled by Host environment. The default process is implemented in DefaultBootstrapper. The runtime processes service assemblies serialized using SCA Assembly XML which can take other forms. - The load phase processes SCDL and creates an in-memory model and produces corresponding runtime artifacts (e.g. components, services, references)
- The connect/wire phase wires references to services
Image Added Anchor |
---|
| Assembly Model |
---|
| Assembly Model |
---|
| Assembly ModelThe SCA assembly model is represented as a set of interfaces in Tuscany. The following are some key elements. - SCA components are configured instances of SCA implementations, which provide and consume services.
- SCA services are used to declare the externally accessible services of an implementation.
- SCA references represent a dependency that an implementation has on a service that is supplied by some other implementation, where the service to be used is specified through configuration.
- An implementation is concept that is used to describe a piece of software technology such as a Java class, BPEL process, XSLT transform, or C++ class that is used to implement one or more services in a service-oriented application. An SCA composite is also an implementation.
- ComponentType refers to the configurable aspects of an implementation.
- Interfaces define one or more business functions. These business functions are provided by Services and are used by components through References. Services are defined by the Interface they implement. SCA currently supports two interface type systems:
- Java interfaces
- WSDL portTypes
- An SCA composite is the basic unit of composition within an SCA Domain. An SCA Composite is an assembly of Components, Services, References, and the Wires that interconnect them.
- SCA wires connect service references to services.
- Bindings are used by services and references. References use bindings to describe the access mechanism used to call the service to which they are wired. Services use bindings to describe the access mechanism(s) that clients should use to call the service.
- Properties allow for the configuration of an implementation with externally set data values. The data value is provided through a Component, possibly sourced from the property of a containing composite.
- Image Added
ContributionThe Tuscany runtime provides a framework to support SCA contributions. The framework can be extended against the following two extension points: PackageProcessorExtensionPoint: It accepts extensions that can handle different packaging format/archives such as a directory, a JAR, an OSGi bundle, an EAR, a WAR and a ZIP. ArtifactProcessorExtensionPoint: It accepts extensions that can handle specific types of artifacts such as WSDL, XSD, composite, java class, BPEL. Image Added - Package processors scans the contribution being installed and generate a list of artifacts that needs to be processed. Currently there is support for folder/file system and jar contribution packages. In order to be available to the contribution service, a package processors needs to register itself with the package processor extension.
- Artifact processors are used to process each artifact available on the contribution. In order to be available to the contribution service, a artifact processor needs to register itself with the artifact processor extension. An artifact processor will be called for each artifact in two phases :
- read phase : This is where you read an artifact (a document, an XML element, classes etc.), populate a model representing the artifact and return it. The SCA contribution service calls ArtifactProcessor.read() on all artifacts that have an ArtifactProcessor registered for them. If your model points to other models, instead of trying to load these other models right away, you should just keep the information representing that reference, which you'll turn into a pointer to the referenced model in the resolve() phase. Note that you don't necessarily need to fully read and populate your model at this point, you can choose to complete it later.
- resolve phase : This phase gives you the opportunity to resolve references to other models (a WSDL, a class, another composite, a componentType). At this point, all models representing the artifacts in your SCA contribution have been read, registered in the contribution's ArtifactResolver, and are ready to be resolved.
- All deployable composites should be now ready to get deployed to the SCA Domain
Anchor |
---|
| Implementation Extension |
---|
| Implementation Extension |
---|
| Implementation ExtensionImplementation extension is responsible for supporting an implementation type, such as Java, Script, and BPEL Anchor |
---|
| Binding Extension |
---|
| Binding Extension |
---|
| Binding ExtensionBinding extension is responsible for supporting a binding type such as Web Service, JMS, JSON-RPC and RMI Anchor |
---|
| Interface Extension |
---|
| Interface Extension |
---|
| Interface ExtensionInterface extension is responsible for supporting an interface type such as Java interface and WSDL 1.1 portType Anchor |
---|
| Data Binding Extension |
---|
| Data Binding Extension |
---|
| Data Binding ExtensionPlease refer to: Tuscany Databinding Guide Anchor |
---|
| Composite Activation |
---|
| Composite Activation |
---|
| Composite ActivationAfter a composite is fully configured, it can be activated in the SCA domain. The Tuscany runtime activates a composite in the following steps: - Build the composite: In this phase, the composite model is further normalized to facilitate the runtime interactions. The metadata is consolidated following the service/reference promotions. With the flattened model, we can get all the information at component level.
- Configure the composite: In this phase, the composite hierarchy is navigated to configure the component implementation, reference binding and service binding with provider factories which will be used in later steps to create runtime wires among components and external services.
- Create the runtime wires: In this phase, runtime wires are created for component references and component services over selected bindings. The runtime wire is a collection of invocation chains that are patitioned by operations. Each invocation chain consists of a set of invokers and interceptors s. Invokers provides the invocation logic to binding protocols and implementation technologies. Interceptors are speical kind of invokers that provide additional logic for the invocation, such as data transformation and transaction control. For a component reference, we create runtime wires to represent the outbound invocation over the selected binding. For a component service, we create runtime wires to represent the inbound invocation to the implementation type. Callback wires can be attached to component services to represent the callback invocations from the service.
- Start the composite: In this phase, the start() callback methods defined by the ImplementationProvider, ReferenceBindingProvider and ServiceBindingProvider will be invoked. As a result, components, component references and component services are initialized to serve the component interactions. Service listeners are started to accept inbound requests from the binding layer.Image Added
Image Added
Invocation Overview- An invocation is dispatched to the WireInvocationHandler
- The WireInvocationHandler looks up the correct InvocationChain
- It then creates a message, sets the payload, sets the TargetInvoker, and passes the message down the chain
- When the message reaches the end of the chain, the TargetInvoker is called, which in turn is responsible for dispatching to the target
- Having the TargetInvoker stored on the outbound side allows it to cache the target instance when the wire source has a scope of equal or lesser value than the target (e.g. request-->composite
The runtime provides components with InboundWires and OutboundWires. InvocationChains are held in component wires and are therefore stateless which allows for dynamic behavior such as introduction of new interceptors or re-wiring. Anchor |
---|
| Loading SCA assemblies |
---|
| Loading SCA assemblies |
---|
| Loading SCA assembliesArtifact ProcessorArtifact processors are responsible for processing each artifact available in a contribution package, these processors are going to be invoked in two phases - read phase : This is where you read an artifact (a document, an XML element, classes etc.), populate a model representing the artifact and return it. The SCA contribution service calls ArtifactProcessor.read() on all artifacts that have an ArtifactProcessor registered for them. If your model points to other models, instead of trying to load these other models right away, you should just keep the information representing that reference, which you'll turn into a pointer to the referenced model in the resolve() phase. Note that you don't necessarily need to fully read and populate your model at this point, you can choose to complete it later.
- resolve phase : This phase gives you the opportunity to resolve references to other models (a WSDL, a class, another composite, a componentType). At this point, all models representing the artifacts in your SCA contribution have been read, registered in the contribution's ArtifactResolver, and are ready to be resolved.
Loading Java SCASCA service assemblies are deployed to the SCA domain in the format of SCDL files. Tuscany runtime artifact processor loads the SCDLs into model objects which are a set of java beans to hold the metadata. There are two types of loaders: - StAXElementLoader: Load the XML element from the StAX events
- ComponentTypeLoader: Load the Component Type for an implementation either by introspection or paring a side file
Loading Component TypeLoads the component type definition for a specific implementation - How it does this is implementation-specific
- May load XML sidefile (location set by implementation)
- May introspect an implementation artifact (e.g. Java annotations)
- ... or anything else
Loading Composite ComponentType Loader - Load SCDL from supplied URL
- Extract and load SCDL from composite package
POJO ComponentType Loader - Introspect Java annotations
- Uses a pluggable "annotation processing" framework to introspect Java classes
Class diagram for the runtime artifactsImage Added Anchor |
---|
| Spring Integration |
---|
| Spring Integration |
---|
| Spring Integration Wiki Markup |
*This page is work in progress. Thanks for your contribution* :-)
* [Overview|#Overview]
** [Core|#Core Definition]
** [Extension|#Extension Definition]
** [Runtime|#Runtime Definition]
* [Internals High Level View|#Internals High Level View]
* [Bootstrap|#Bootstrap]
* [Metadata|#Metadata]
* [Contribution|#Contribution]
* [Binding Extension|#Binding Extension]
* [Component Implementation Extension|#Component Implementation Extension]
* [Data Binding Extension|#Data Binding Extension]
* [Wiring|#Wiring]
* [Loading SCA assemblies|#Loading SCA assemblies]
* [Spring Integration|#Spring Integration]
** [Spring as component implementation|#Spring as component implementation]
** [Spring as IOC container|#Spring as IOC container]
h3. {anchor:Overview} {color:#0099cc}Overview{color}
The SCA Java runtime is composed of core and extensions. The core is essentially a multi-VM wiring engine that connects components together using the principles of [Dependency Injection|http://en.wikipedia.org/wiki/Dependency_injection], or [Inversion of Control|http://en.wikipedia.org/wiki/Inversion_of_control].
h4. {anchor:Core Definition} Core
The Core is designed to be simple and limited in its capabilities. It wires functional units together and provides SPIs that extensions can interact with. Capabilities such as service discovery, reliability, support for transport protocols, etc. are provided through extensions.
h4. {anchor:Extension Definition} Extension
Extensions enhance SCA runtime functionality. Extension types are not fixed and core is designed to be as flexible as possible by providing an open-ended extension model. However, there are a number of known extension types defined including:
* *Component implementation types*, e.g. Spring, Groovy, and JavaScript
* *Binding types*, e.g. Axis, CXF, AMQP, ActiveMQ, JXTA
* *DataBinding types*, e.g. JAXB, SDO, XmlBeans
* *Interface Binding types*, e.g. WSDL, Java
Details of how to implement an extension can be can be found in the [Extensions Guide| Java SCA Extension Developer Guide].
h4. {anchor:Runtime Definition} Runtime
The core is designed to be embedded in, or provisioned to, a number of different host environments. For example, the core may be provisioned to an OSGi container, a standalone runtime, a servlet engine, or J2EE application server. Runtime capabilities may vary based on the host environment.
{HTMLcomment:hidden}{children:sort=creation}{HTMLcomment}
h4. High level overview of Java SCA runtime
The diagram shown below is a high level view of the SCA runtime which consists of the following key modules/packages:
# SCA Spec API: The APIs defined by the SCA Java Client and Implementation Spec
# API: Tuscany APIs which extend the SCA spec APIs
# Core: The runtime implementation and the SPIs to extend it
# Extensions:
## Container implementation - For extending language support, for example BPEL, Python, C++, Ruby,..
## Binding - for extending protocol support, for example Axis2, CXF,..
## Interface Binding - for service definition, for example WSDL, Java, ...
## Databinding - for extending data support, for example SDO, JAXB, ...
# Host Platforms: The environments that host the Tuscany runtime.
!architectureOverview.jpg!
h3. {anchor:Internals High Level View}{color:#0099cc}Internals High Level View{color}
\[Note: Do we want to link to [Kernel-Structure|Java SCA Modulization Design Discussions]\]?
h3. {anchor:Bootstrap}{color:#0099cc}Bootsrap{color}
Bootstrap process is controlled by Host environment. The default process is implemented in DefaultBootstrapper.
The runtime processes service assemblies serialized using SCA Assmebly XML which can take other forms.
* The load phase processes SCDL and creates an in-memory model and produces corresponding runtime artifacts (e.g. components, services, references)
* The connect/wire phase wires references to services
!bootstrap.jpg!
h3. {anchor:Metadata}{color:#0099cc}Metadata{color}
!assembly.jpg!
h3. {anchor:Contribution}{color:#0099cc}Contribution{color}
!contribution_overview.jpg!
* Package processors scans the contribution being installed and generate a list of artifacts that needs to be processed. Currently there is support for folder/file system and jar contribution packages. In order to be available to the contribution service, a package processors needs to register itself with the package processor extension.
* Artifact processors are used to process each artifact available on the contribution. In order to be available to the contribution service, a artifact processor needs to register itself with the artifact processor extension. An artifact processor will be called for each artifact in two phases :
** read phase : This is where you read an artifact (a document, an XML element, classes etc.), populate a model representing the artifact and return it. The SCA contribution service calls ArtifactProcessor.read() on all artifacts that have an ArtifactProcessor registered for them. If your model points to other models, instead of trying to load these other models right away, you should just keep the information representing that reference, which you'll turn into a pointer to the referenced model in the resolve() phase. Note that you don't necessarily need to fully read and populate your model at this point, you can choose to complete it later.
** resolve phase : This phase gives you the opportunity to resolve references to other models (a WSDL, a class, another composite, a componentType). At this point, all models representing the artifacts in your SCA contribution have been read, registered in the contribution's ArtifactResolver, and are ready to be resolved.
* All deployable composites should be now ready to get deployed to the SCA Domain
h3. {anchor:Binding Extension}{color:#0099cc}Binding Extension{color}
h3. {anchor:Component Implementation Extension}{color:#0099cc}Component Implementation Extension{color}
h3. {anchor:Data Binding Extension}{color:#0099cc}Data Binding Extension{color}
h3. {anchor:Wiring}{color:#0099cc}Wiring{color}
[This section needs to be updated]
To understand how wiring works, we need to detail how Components function in the system. Let's start with Atomic Components and then discuss Composite Components.!runtime-wire.jpg!
h4. Atomic Component
AtomicComponent is the most basic component form. It corresponds to the spec concept which offers services, has references and properties
* Implementation types e.g. Java, XSLT, etc.
* Are wired
* Have properties
Atomic Components have implementation instances. Instances are associated with a scope: e.g. request, conversation, composite. A SCDL entry is used to define a Component.
Atomic components use a ScopeContainer to manage implementation instances:
* Composite, HTTP Session, Request, Stateless
* ScopeContainers track implementation instances by scope id and the AtomicComponent instance identity
* Instances are stored in an InstanceWrapper which is specific to the component implementation type (e.g. PojoInstanceWrapper.java)
h4. Component Wiring
Component references are connected to services through wires
* Two sides
** InboundWire - handles the source side of a wire, including policy
** OutboundWire - handles the target side of a wire, including policy
* The runtime connects inbound and outbound wires, performing optimizations if possible
** Inbound and outbound wires may be associated with different service contracts
** Different implementation types
** "Standard" wires contain invocation chains that have Interceptors that perform some form of mediation (e.g. policy)
** Other wire types exist that, for example, do not perform mediations
h4. Invocation Chains
(?) is "around-style" mediation an accepted term. If so what does it mean. Googling it doesn't return anything appropriate.
A wire has an InvocationChain per service operation. An InvocationChain may have interceptors - "around-style" mediation. Component implementation instances access a wire through a WireInvocationHandler associated with a reference.
* WireInvocationHandlers may (or may not depending on the component type) be fronted by a proxy
* WireInvocationHandlers dispatch an invocation to the correct chain
A wire has a TargetInvoker that is created from the target side AtomicComponent or Reference and is stored on the source wire invocation handler. The TargetInvoker is resposnible for dispatching the request to the target instance when the message hits the end of the target invocation chain.
h4. Invocation Overview
* An invocation is dispatched to the WireInvocationHandler
* The WireInvocationHandler looks up the correct InvocationChain
* It then creates a message, sets the payload, sets the TargetInvoker, and passes the message down the chain
* When the message reaches the end of the chain, the TargetInvoker is called, which in turn is responsible for dispatching to the target
* Having the TargetInvoker stored on the outbound side allows it to cache the target instance when the wire source has a scope of equal or lesser value than the target (e.g. request-->composite
The runtime provides components with InboundWires and OutboundWires. InvocationChains are held in component wires and are therefore stateless which allows for dynamic behavior such as introduction of new interceptors or re-wiring.
h3. {anchor:Loading SCA assemblies }{color:#0099cc}Loading SCA assemblies{color}
h4. Artifact Processor
Artifact processors are responsible for processing each artifact available in a contribution package, these processors are going to be invoked in two phases
* read phase : This is where you read an artifact (a document, an XML element, classes etc.), populate a model representing the artifact and return it. The SCA contribution service calls ArtifactProcessor.read() on all artifacts that have an ArtifactProcessor registered for them. If your model points to other models, instead of trying to load these other models right away, you should just keep the information representing that reference, which you'll turn into a pointer to the referenced model in the resolve() phase. Note that you don't necessarily need to fully read and populate your model at this point, you can choose to complete it later.
* resolve phase : This phase gives you the opportunity to resolve references to other models (a WSDL, a class, another composite, a componentType). At this point, all models representing the artifacts in your SCA contribution have been read, registered in the contribution's ArtifactResolver, and are ready to be resolved.
h3. Loading Java SCA
SCA service assemblies are deployed to the SCA domain in the format of SCDL files. Tuscany runtime artifact processor loads the SCDLs into model objects which are a set of java beans to hold the metadata.
There are two types of loaders:
* StAXElementLoader: Load the XML element from the StAX events
* ComponentTypeLoader: Load the Component Type for an implementation either by introspection or paring a side file
h3. Loading Component Type
Loads the component type definition for a specific implementation
* How it does this is implementation-specific
* May load XML sidefile (location set by implementation)
* May introspect an implementation artifact (e.g. Java annotations)
* ... or anything else
Loading Composite ComponentType Loader
* Load SCDL from supplied URL
* Extract and load SCDL from composite package
POJO ComponentType Loader
* Introspect Java annotations
* Uses a pluggable "annotation processing" framework to introspect Java classes
h3. Class diagram for the runtime artifacts
!runtime-model.jpg!
h3. {anchor:Spring Integration}{color:#0099cc}Spring Integration{color}
|