The Clacks Project aims to provide a general infrastructure service framework which can be used for various tasks. It is written in Python and makes use of AMQP messaging for reliable communication.
Clacks was designed as a backend for the GOsa GUI, but does not depend on it. It has not been named GOsa-whatever to avoid naming confusion.
The AMQP infrastructure is divided into domains. This makes it possible to separate different locations by administration logic and/or access control. Clients can be “joined” to the domain.
Agent infrastructure with varying plugin set
Multiple agents can connect to the AMQP infrastructure of your domain but do not necessarily need to provide the same functionality. Message routing will push the message to the correct agent.
Automatic load balancing
Plugins provided by more than one agent automatically benefit from automatic load balancing.
Additionally to agents, clients can be registered/joined to your domain. You can simply provide new client functionality by a DBUS/shell proxy. Your clients get an automatic inventory which can be stored by an agent interested in it.
Easy to register RPC methods
From the programmers point of view, plugins just need to inherit the Plugin class and export their methods using the @Command decorator. This is unique for clients and agents.
The RPC caller does not need to know where the called functionality is provided. It automatically reaches the correct location.
XML based eventing system
Components can send validated XML events which can be interpreted by other components. I.e. an asterisk VoIP call can be signalized on that event bus, and a statistics module can subscribe for these events to generate its statistics.
Object abstraction for various backends
The Clacks framework comes with a XML Xquery engine that allows to map certain information to various backends. If you - for example - need additional attributes for the predefined “User” object, you can simply add them to the object description and “wire” it to target attributes in your LDAP or relational database.
SASL and SSL support
Shell and scripting
Every exported method can be easily used inside the Clacks shell or from simple Python scripts. Other languages can be used, too. They just need to be able to speak JSONRPC over HTTP[S] or AMQP[S].