Apologies if this overlaps with a previously posted question. I've been using an RStudio Server Professional license for a few years to run a smallish server for the use of my honours students, especially as they do senior courses in statistics. RStudio was kind enough to make this available to me.
I'm finding myself needing to ramp up the number of students locally on campus who access this resource, and need an idea of what kind of resources I really should have available. I found the recommendations/FAQ, which suggests that for 15 analysts with a mix of job loads, 16 cores and 256GB of RAM is recommended. I would be looking at more like 50-60 concurrent users, but with minimal job loads on the whole.
So, assuming I did get 2 or more physical boxes, probably dual-processor Xeons (say, 2 10-core Xeons, for sake of argument), similar to a box I already have that runs other educational products, would this suffice? Assume 2 or 3 HP servers, 20 cores each, 128GB+ RAM each. Some minimal disk, and then all user accounts mounted across the network.
Just looking for anyone who's actually implemented such a system, and their experiences. If you run a corporate system with low-workload analysts regularly, I'd also love to hear your experiences. I need to build a budget for the incredibly arcane process of academic funding, so I need a ballpark for total computational hardware before I even begin.