Using Rhino/GH to create a webUI parametric modeling platform.

Hi 

I'm in the early stages of attempting to develop a parametric modeling tool which would streamline one of my clients online shop.

The site I'm working on is for an engineer who build custom trays and tool boxes for ute/pickups. It would be really awesome if we could work out a way for the customers to input measurements, model of vehicle, layout etc. Then have a 3D model on screen to display the product. There would probably be a selection of designs styles and material finished for each element then they would choose a layout (in a grid, almost like laying out a site). The end game would also have the ability to flatten/unfold some elements so they can be sent directly to the cutting machines. Produce schedules for ordering, pricing, folding and the works.

I have been searching around and people have advised me that Rhino/GH wont be easy to use for a web service and I would be better off building a JS/GL application from scratch. Like these guys have done. https://n-e-r-v-o-u-s.com/cellCycle/ You can actually access all the JS files to have a look at how it works.

However, I think that's a bit over kill. I am hoping there may be a way to retool existing software into a workflow which uses Rhino/GH also this would keep the software in my comfort zone and offer a platform that can be easily updated and modified.

What I am proposing is a way to provision a pre-setup and suspended VM for each instance of end user interaction. I understand that Rhino is able to work with C+

- Web UI populates a database of the required values.
- The user hits the "build" button.
- Server provisions (maybe clone the process? to avoid load time? that's where I was going with the VM's) a new instance of rhino/GH.
- A script tells the instance where to load the values from.
- Rhino/GH produces model, bakes and exports it.
- Server then grabs the exported file and uploads it into a webGL viewer.
- Process repeats each time user makes a change.  

There will be a lot of time spent refining the GH definitions and they dont need to be hugely complex o lower the processing time on each one however, I intend to run this on a top of the line server with SSDs so hopefully all the above can be executed in a matter of seconds. File size and UL/DL speeds will be the bottle neck. This also means the server could probably handle several simultaneous instances. Due to existing traffic this should be manageable. 

Does this sound possible? Does anyone have any experience or insights into this?

Cheers guys!