Deployment

A deployment is the process of initialising an architecture, creating threads corresponding to components, and synchronizing and strating these threads.

A Deployment is then the main entry point of your MAUVE application. You can create a deployment either as a plain C++ executable, or as a library that will be loaded using the Scala Deployer.

Deployment Processes

The following code defines a deployment process that can be built as an executable:

int main(int argc, char** argv) {
  int r = ::mauve::runtime::main(argc, argv);
  if (r > 0) return r;

  auto archi = new MyArchitecture();
  auto depl = mk_abstract_deployer(archi);
  archi->configure();
  depl->create_tasks();
  depl->activate();
  depl->start();
  depl->loop();
  depl->stop();
  archi->cleanup();
  return 0;
}

The first line calls an helper function, mauve::runtime::main, that provides options to load configuration files (logging and property configurations). Then an architecture objet archi is created, and bind to a deployer depl using mk_abstract_deployer. The deployement steps are then to:

• configure the architecture (archi->configure())
• create component tasks (depl->create_tasks())
• activate component tasks (depl->activate())
• start the component tasks (depl->start())
• suspend the deployer execution while the architecture components execute (depl->loop())

If the process is terminated (using a signal for instance), then steps to cleanly end the deployment are:

• stopping the component tasks (depl->stop())
• cleaning the architecture (archi->cleanup())

Scala Deployment

Documentation In Progress