Noticed Aaron’s post recently about some old Apollo DomainOS workstations. If I remember right, Apollo was purchased by Hewlett Packard before the HPUX OS was created. Without researching it, I’d say this would have been somewhere in the late 80’s to early 90’s.
Anyway, the DomainOS system had something interesting – a distributed file system. File systems could be cross-linked between different workstations back and forth. Changing directory to // and doing an ls would show all workstations in the group.
I’d have to say it was ahead of it’s time, as the next thing to come down the pipe was NFS. Anyone who had much experience at the time with NFS and cross-mounting between workstations has probably seen what could happen to single-threaded NFS processes when something caused a hiccup on the network. NFS timeouts would roll through the network, and if it got bad enough, local hard drives would even stop responding. There were a few instances in the network I was familiar with where all the HPUX workstations had to be shut down and started back up in a certain sequence to break the NFS timeout loops. IT guys coming from DomainOS with it’s distributed file system soon learned the hard way about centralized file storage and the no-no of having a NFS client also be a NFS server.