Local cloud, or LAN, Combining CUDA Processing from multiple CUDA Desktops

I’d like to utilize the Titan RTX’s off of multiple desktops , xth or as many desktops as I want via local area network. What’s the best approach? In other words, I don’t want to chunk together one large SLI’d multiple GPU Server, I’d like to utilize the desktop workstations CUDA cores as well as a server. Thanks!