There are no official tools to create a big Swarm at a scale yet. At the moment, operators use internal scripts, ad hoc tools (such as Belt), configuration managers (such as Puppet or Ansible), or orchestration templates (such as CloudFormation for AWS or Heat for OpenStack), as we have seen in the previous chapters. Recently, Docker For AWS and Azure came as alternatives.

But this use case will be probably addressed in a unified way with software-defined infrastructure toolkits.