{section}
{column}
h1. On the Petals ESB Topology
Petals ESB is a fully distributed ESB that features many types of topology both dynamically or statically. The topology of a Petals ESB cluster can be configured through a topology definition (ie. {{topology.xml}}) and [server properties|Container Configuration] (ie. {{server.properties}}).
This document will explain to you how to properly configure the topology of your Petals ESB cluster.
{column}
{column:width=350px}
{panel:title=Table of contents}{toc}{panel}
{panel:title=Contributors}{contributors:order=name|mode=list}{panel}
{column}
{section}
h1. Topology composition
A Petals ESB cluster is composed of a *domain* divided into *sub-domains* that contain the *containers*. A sub-domain aims to limit the visibility of service endpoints.
Inside a sub-domain:
* all service providers can be seen by all service consumers without restrictions,
* all service providers must be accessible directly to all service consumers.
Using a gateway, it is possible to interact between sub-domains pairs:
* by default, service endpoints of a sub-domain are not propagated to other sub-domains,
* the gateway is configured to propagated a selection of service endpoints in both ways,
* a message sent from a service consumer located into the sub-domain #1 to a service provider located in sub-domain #2 will pass through the gateway.
*Note: in the current version of Petals ESB, the notion "sub-domain" is not fully implemented, and the "gateway" is not yet available. The processing is as if the domain contains only one sub-domain: all service endpoints can be seen by all.*
{gliffy:name=topology-subdomain|version=5|size=L|align=center}
h1. Configuring the sub-domain
The sub-domain configuration is composed of several part:
* the sub-domain identification,
* the sub-domain mode,
* the sub-domain registry configuration.
h2. Configuring the sub-domain mode
A Petals sub-domain supports two different modes :
* static : no new container can be added to a running Petals sub-domain. All containers of the sub-domain must be stop and restart to take into account changes
* dynamic : a container can be attached to or detached from a running sub-domain, without requiring a full start/stop of the sub-domain
h3. Choosing the sub-domain mode
The sub-domain mode 'static' is recommended in following use-case:
* you want to prevent to add or remove container from your sub-domain,
* you don't plan hot topology refactoring
If you have a worst constraint about the availability of the Petals ESB (24/24, 7/7), you shoudl use the mode 'dynamic' to be able to apply topology changes without stopping the Petals ESB.
h3. Using the static mode
To use the static mode, in the sub-domain element of the topology definition, you just have to change the mode property to *static* :
{code:lang=xml}<tns:sub-domain name="your-subdomain-name" mode="static">{code}
See : [^static-topology-sample.xml]
h3. Using the dynamic mode
To use the dynamic mode, in the sub-domain element of the topology definition, you just have to change the mode property to *dynamic* :
{code:lang=xml}<tns:sub-domain name="your-sub-domain-name" mode="dynamic">{code}
See : [^dynamic-topology-sample.xml]
h2. Configuring the sub-domain registry
h3. Configuring the implementation to use
Firstly, the implementation of the sub-domain registry must be defined through the tag '{{registry-implementation}}' that contains the class name of the implementation to use:
{code}
<tns:sub-domain name="subdomain-0" mode="static" xmlns:tns="http://petals.ow2.org/topology">
...
<tns:registry-implementation>org.ow2.petals.microkernel.registry.overlay.RegistryOverlayImpl</tns:registry-implementation>
...
</tns:sub-domain>
{code}
The default implementation is the Petals Registry Overlay is used. And the tag '{{registry-implementation}}' cab be omitted.
h3. Configuring the implementation itself
Secondly, the implementation itself must be configured using an extension of the sub-domain tag. See the [documentation of the registry implementation|Petals ESB "Registry"#Implementations] for more information.
h1. Declaring containers configurations
As we have seen, the {{topology.xml}} resource is used to declare the configuration of a Petals ESB cluster.
h2. Generic container configuration
Each Petals ESB container declared within must conforms to the following XML structure :
{code:lang=xml}
<tns:container name="0">
<tns:description>description of the container 0</tns:description>
<tns:host>localhost</tns:host>
<tns:user>petals</tns:user>
<tns:password>petals</tns:password>
<tns:webservice-service>
<tns:port>7600</tns:port>
<tns:prefix>petals/ws</tns:prefix>
</tns:webservice-service>
<tns:jmx-service>
<tns:rmi-port>7700</tns:rmi-port>
</tns:jmx-service>
<tns:transport-service>
<tns:tcp-port>7800</tns:tcp-port>
</tns:transport-service>
</tns:container>
{code}
The meaning of these properties is summarized in the following table :
{center}
|| Property name || Default value || Required || Description ||
| {{name}} | 0 | yes | name of the container |
| {{description}} | description of the container 0 | no | description of the container |
| {{host}} | localhost | yes | host name or ip address of the container |
| {{user}} | petals | no | jmx username |
| {{password}} | petals | no | jmx password |
|| {{webservice-service}} || || || ||
| {{port}} | 7600 | yes | administration webservices port (topology, information, ...) |
| {{prefix}} | petals/ws | yes | path of the webservice |
|| {{jmx-service}} || || || ||
| {{rmi-port}} | 7700 | yes | jmx service port |
|| {{transport-service}} || || || ||
| {{tcp-port}} | 7800 | yes |NIO transporter port |
{center}
h1. Configuring the local container
As we have seen, the {{topology.xml}} file contains the declaration of a Petals ESB cluster. This information is not sufficient to determine which configuration must be chosen by the local container at startup. To do so, you must give a container name declared within the {{topology.xml}} file to the property {{petals.container.name}} declared in the {{server.properties}}.
{noformat}
#This property specifies the name of the container. In distributed mode, this property is mandatory
#and must match a container name in the topology.xml file
petals.container.name=0
{noformat}
The container name given to the {{petals.container.name}} property must be declared within the {{topology.xml}} file, otherwise, Petals ESB won't be able to start. This mean that you must have declared the local container configuration within the {{topology.xml}} file.
h1. Appendix
h2. Topology XML Schema
For more information on the structure of the topology.xml file, you can consult its XML schema definition here : [^petalsTopology.xsd].
{column}
h1. On the Petals ESB Topology
Petals ESB is a fully distributed ESB that features many types of topology both dynamically or statically. The topology of a Petals ESB cluster can be configured through a topology definition (ie. {{topology.xml}}) and [server properties|Container Configuration] (ie. {{server.properties}}).
This document will explain to you how to properly configure the topology of your Petals ESB cluster.
{column}
{column:width=350px}
{panel:title=Table of contents}{toc}{panel}
{panel:title=Contributors}{contributors:order=name|mode=list}{panel}
{column}
{section}
h1. Topology composition
A Petals ESB cluster is composed of a *domain* divided into *sub-domains* that contain the *containers*. A sub-domain aims to limit the visibility of service endpoints.
Inside a sub-domain:
* all service providers can be seen by all service consumers without restrictions,
* all service providers must be accessible directly to all service consumers.
Using a gateway, it is possible to interact between sub-domains pairs:
* by default, service endpoints of a sub-domain are not propagated to other sub-domains,
* the gateway is configured to propagated a selection of service endpoints in both ways,
* a message sent from a service consumer located into the sub-domain #1 to a service provider located in sub-domain #2 will pass through the gateway.
*Note: in the current version of Petals ESB, the notion "sub-domain" is not fully implemented, and the "gateway" is not yet available. The processing is as if the domain contains only one sub-domain: all service endpoints can be seen by all.*
{gliffy:name=topology-subdomain|version=5|size=L|align=center}
h1. Configuring the sub-domain
The sub-domain configuration is composed of several part:
* the sub-domain identification,
* the sub-domain mode,
* the sub-domain registry configuration.
h2. Configuring the sub-domain mode
A Petals sub-domain supports two different modes :
* static : no new container can be added to a running Petals sub-domain. All containers of the sub-domain must be stop and restart to take into account changes
* dynamic : a container can be attached to or detached from a running sub-domain, without requiring a full start/stop of the sub-domain
h3. Choosing the sub-domain mode
The sub-domain mode 'static' is recommended in following use-case:
* you want to prevent to add or remove container from your sub-domain,
* you don't plan hot topology refactoring
If you have a worst constraint about the availability of the Petals ESB (24/24, 7/7), you shoudl use the mode 'dynamic' to be able to apply topology changes without stopping the Petals ESB.
h3. Using the static mode
To use the static mode, in the sub-domain element of the topology definition, you just have to change the mode property to *static* :
{code:lang=xml}<tns:sub-domain name="your-subdomain-name" mode="static">{code}
See : [^static-topology-sample.xml]
h3. Using the dynamic mode
To use the dynamic mode, in the sub-domain element of the topology definition, you just have to change the mode property to *dynamic* :
{code:lang=xml}<tns:sub-domain name="your-sub-domain-name" mode="dynamic">{code}
See : [^dynamic-topology-sample.xml]
h2. Configuring the sub-domain registry
h3. Configuring the implementation to use
Firstly, the implementation of the sub-domain registry must be defined through the tag '{{registry-implementation}}' that contains the class name of the implementation to use:
{code}
<tns:sub-domain name="subdomain-0" mode="static" xmlns:tns="http://petals.ow2.org/topology">
...
<tns:registry-implementation>org.ow2.petals.microkernel.registry.overlay.RegistryOverlayImpl</tns:registry-implementation>
...
</tns:sub-domain>
{code}
The default implementation is the Petals Registry Overlay is used. And the tag '{{registry-implementation}}' cab be omitted.
h3. Configuring the implementation itself
Secondly, the implementation itself must be configured using an extension of the sub-domain tag. See the [documentation of the registry implementation|Petals ESB "Registry"#Implementations] for more information.
h1. Declaring containers configurations
As we have seen, the {{topology.xml}} resource is used to declare the configuration of a Petals ESB cluster.
h2. Generic container configuration
Each Petals ESB container declared within must conforms to the following XML structure :
{code:lang=xml}
<tns:container name="0">
<tns:description>description of the container 0</tns:description>
<tns:host>localhost</tns:host>
<tns:user>petals</tns:user>
<tns:password>petals</tns:password>
<tns:webservice-service>
<tns:port>7600</tns:port>
<tns:prefix>petals/ws</tns:prefix>
</tns:webservice-service>
<tns:jmx-service>
<tns:rmi-port>7700</tns:rmi-port>
</tns:jmx-service>
<tns:transport-service>
<tns:tcp-port>7800</tns:tcp-port>
</tns:transport-service>
</tns:container>
{code}
The meaning of these properties is summarized in the following table :
{center}
|| Property name || Default value || Required || Description ||
| {{name}} | 0 | yes | name of the container |
| {{description}} | description of the container 0 | no | description of the container |
| {{host}} | localhost | yes | host name or ip address of the container |
| {{user}} | petals | no | jmx username |
| {{password}} | petals | no | jmx password |
|| {{webservice-service}} || || || ||
| {{port}} | 7600 | yes | administration webservices port (topology, information, ...) |
| {{prefix}} | petals/ws | yes | path of the webservice |
|| {{jmx-service}} || || || ||
| {{rmi-port}} | 7700 | yes | jmx service port |
|| {{transport-service}} || || || ||
| {{tcp-port}} | 7800 | yes |NIO transporter port |
{center}
h1. Configuring the local container
As we have seen, the {{topology.xml}} file contains the declaration of a Petals ESB cluster. This information is not sufficient to determine which configuration must be chosen by the local container at startup. To do so, you must give a container name declared within the {{topology.xml}} file to the property {{petals.container.name}} declared in the {{server.properties}}.
{noformat}
#This property specifies the name of the container. In distributed mode, this property is mandatory
#and must match a container name in the topology.xml file
petals.container.name=0
{noformat}
The container name given to the {{petals.container.name}} property must be declared within the {{topology.xml}} file, otherwise, Petals ESB won't be able to start. This mean that you must have declared the local container configuration within the {{topology.xml}} file.
h1. Appendix
h2. Topology XML Schema
For more information on the structure of the topology.xml file, you can consult its XML schema definition here : [^petalsTopology.xsd].