As an ex VFX sysadmin, this is a little underwhelming.
its trivial to provide a machine on demand, We used to do it all the time. Because the version of software are known in advance[1], its super simple to make sure they are there and ready for the user. Our machine images were like 120 gigs, with 3/4 of the software on NFS shares. (still much faster than running in docker containers)
What they dont cover is the actual important bits:
o keeping the session colour correct
o making it responsive.
Remote workstations have been about for many years. Mainly because the big workstations were too noisy to use with clients. so they used teradici cards to allow dual screen 2k screens.
The third thing they don't really cover, but I suspect thats out of scope is how they sync assets. VFX is still a file based system, which translates poorly to object storage, even worse to object storage that doesn't support seek()
[1] vfx pipelines used fixed versions of software, because otherwise everything turns to shit.
its trivial to provide a machine on demand, We used to do it all the time. Because the version of software are known in advance[1], its super simple to make sure they are there and ready for the user. Our machine images were like 120 gigs, with 3/4 of the software on NFS shares. (still much faster than running in docker containers)
What they dont cover is the actual important bits:
o keeping the session colour correct
o making it responsive.
Remote workstations have been about for many years. Mainly because the big workstations were too noisy to use with clients. so they used teradici cards to allow dual screen 2k screens.
The third thing they don't really cover, but I suspect thats out of scope is how they sync assets. VFX is still a file based system, which translates poorly to object storage, even worse to object storage that doesn't support seek()
[1] vfx pipelines used fixed versions of software, because otherwise everything turns to shit.