Add flags and inputs to support deploying multiple pairs of the same network on a single machine

[rswanson]

• Needs to support having the instance number (for example --instance 1) as part of the data dir file path.
• Need to either support inputting the instance numbers by hand or auto generating the set of instance numbers. Probably start with the former
• also worth thinking about how these inputs would play if the backend changed from a vanilla server to a k8s cluster. In theory the deployment logic itself changes, but i think conceptually it would something like run N instances of a given network within a namespace as a good parallel

[rswanson]

this probably requires implementing a better (read: in code) means of managing the service files and start scripts as opposed to the current method of copying existing files. What I probably want to do is figure out how to represent each as a template in code and then accept a config object containing the set of things that would exist on the template.

[rswanson]

effectively each *.service needs to be defined as *.N.NETWORK.service meaning that N is an integer that corresponds to either the --instance flag or to the number we add to the set of default ports for each EL/CL client