We are looking into building a solution around the Tesla M6/M60. I have a question regarding the licensing.
Today we are using K1/K2 that connects to(vDGA) a number of virtual machines(VMs) in our vSphere 6.5 environment. The VMs are running RHEL7.4.
The users then connect to the VMs using an application that is basically a VNC connection, multiple users connect to each VM where they run the application that is using the GPU.
In the future we would like to run vGPU so we are looking at GRID2/Tesla for that.
So my first question is: In our setup(see above) what would a CCU be in a license context ? Is it the number of VMs that has GPUs added to the it ? Or is it somehow the number of endusers ?
Second question is: If we buy a M6/M60 but do not use it with vGPU. Instead we use vDGA(passthrough), will i still need a license or is this only when using vGPU ?
Third question: What licensform would be ideal for our setup ? The "GRID Virtual PC" doesnt seem to support over 1GB memory per profile so that will not work. Can someone explain the difference between "GRID Virtual Apps" and "Quadro Virtual Datacenter Workstation". Ive read this http://images.nvidia.com/content/grid/pdf/161207-GRID-Packaging-and-Licensing-Guide.pdf but i dont really understand it so if someone could make it clearer i would appreciate it.