Optimizing simultaneous startup or modification of inter-dependent machines with specified priorities

A computer-implemented method of parallel administration of a multi-machine computing system, the method comprising:

• identifying a plurality of nodes that correspond to individual machines of the multi-machine computing system;
• obtaining, at the at least one of the individual machines, estimated total administration times for each of the nodes;
• obtaining, at the at least one of the individual machines, administration priorities for each of the nodes;
• constructing, in at least one of the individual machines, a graph of dependencies among the nodes;
• identifying, at the at least one of the individual machines, availability of a first set of administration resources to assist in administration of one or more of the individual machines;
• selecting, at the at least one of the individual machines, a first set of the nodes for administration in response to the graph of dependencies, administration priorities, estimated total administration times, and availability of the first set of administration resources;
• administering a first set of machines, corresponding to the first set of the nodes, in parallel using the first set of administration resources;
• updating the graph of dependencies in response to administration of the first set of machines;
• selecting, at the least one of the individual machines, a subsequent set of nodes for administration in response to the updated graph of dependencies, administration priorities, estimated total administration times, and availability of a subsequent set of administration resources;
• and
• administering a subsequent set of machines, corresponding to the subsequent set of the nodes, in parallel using the subsequent set of administration resources.