I imagine what makes this especially hard is you have (at least) three parties in play here: - the people doing the research - the institution's IT services g

I imagine what makes this especially hard is you have (at least) three parties in play here: - the people doing the research - the institution's IT services group - the administrator who writes the checks And in my experience, "actual knowledge of what must be done and what it will or could cost" can vary greatly across these three groups; frequently in very unintuitive ways.