Abstract: A node of a fault-tolerant system relies upon a primary communication mechanism, when possible, for primary connection to a central server. The node monitors its primary connection to the server. While the primary connection is active, the node operates in a supervised mode, in that it generally does not perform a certain subset of tasks without having received real-time commands from the server to perform those tasks. However, when the node detects that it is no longer connected to the server, the node transitions into a fail-over mode. The node operates in the fail-over mode until the node detects that primary connection is re-established. While in fail-over mode, the node may execute a stored set of fail-over instructions that were provided by the server. The node may also attempt to establish a backup connection to the server via a second and different type of communication mechanism.