We have a very precise use case at work.
I work in academic research. So most of our processing of scientific data is done on huge Linux clusters.
But there are a few case where there is a REALLY BIG amount of data to process, and the first step:
- is very cheap to execute
- drastically reduces the amount of data.
So instead of upload several gigabytes of data, let the scientist pre-process the first step on their workstation and only send the smaller amount of already processed data to the cluster for further much more CPU-intensive processing.
Problem: not all biologist run Linux or Mac OS X, and most of our scientific software is designed for Unix (and is a PITA to get to compile on Windows).
Current solution: for windows user, we ship a giant virtual machine that they can run in virtualbox. (Again, the pre-processing isn't CPU intensive, so the VM doesn't impact it that much and is a working solution).
WSL on Windows 10 solution: we could ship the Linux executable, and have Windows 10 users use it to do the pre-processing.
I work in academic research. So most of our processing of scientific data is done on huge Linux clusters.
But there are a few case where there is a REALLY BIG amount of data to process, and the first step:
- is very cheap to execute
- drastically reduces the amount of data.
So instead of upload several gigabytes of data, let the scientist pre-process the first step on their workstation and only send the smaller amount of already processed data to the cluster for further much more CPU-intensive processing.
Problem: not all biologist run Linux or Mac OS X, and most of our scientific software is designed for Unix (and is a PITA to get to compile on Windows).
Current solution: for windows user, we ship a giant virtual machine that they can run in virtualbox. (Again, the pre-processing isn't CPU intensive, so the VM doesn't impact it that much and is a working solution).
WSL on Windows 10 solution: we could ship the Linux executable, and have Windows 10 users use it to do the pre-processing.
Comment