Research is underway on a variety of ion thrusters. For an integrated program, a standard thruster and its interface to the standard power module must be specified. A "unit size" can be specified, with some number of these serving as the thruster with a large power module. In turn arrays of these can be assembled for extremely large payload such as manned Mars freight, with most payloads requiring only one engine, possibly not maximum size. The choice of fuel depends on other considerations besides performance, in particular cost and supply, so that xenon for example might not be the best choice.
The example mission in Some Mars Trajectory Optimizations (PDF) ignores the exhaust velocity. It had been the author's experience that this was typically achievable. For one example, 230 days with 54000 Kg of fuel at 1.15 Mw requires an exhaust velocity of 29000. Values not greatly different from this require not greatly suboptimal times.
The discussion here suggests that SIP yields a robust program. For a wide variety of applications, including interplanetary, IP need only operate near Earth, since aerocapture can be used at target planets.
The major structural components of an ion engine can be recovered after a flight, by allowing a small fraction of the flight for the return. The thrusters can be replaced, and any other refurbishment carried out, at the Earth station.