For customers using CDC software, CDC 180 machines booted into the old 170-mode operating system named NOS. This stood for Network Operating System--somewhat of a misnomer. From NOS, you could then boot CDC's new 180-mode operating system, NOS/VE. VE stood for Virtual Environment; i.e., this OS used virtual memory. In 1985, NOS/VE wasn't really ready for prime time, and I expect quite a few customers simply ran NOS and didn't use the 180 mode of their machines at all. But I think it was possible--in fact, mandatory--to continue running NOS in the background after NOS/VE booted. NOS/VE, at least at this stage in its life (version 1.1.2), required NOS (then at version 2.4.1) to perform some I/O functions. (See footnote.) See the sample login for a user's view of this.
NOS/VE was a totally new operating system, written from the ground up in a new programming language. You've got to admire CDC's gumption. The OS took some inspiration from NOS and from UNIX, which was becoming an important operating system, but in no way was NOS/VE supposed to be a clone of or even compatible with any other OS.
Unlike UNIX, and more like VAX/VMS, VE attempted to be a complex operating system with a lot of features built-in to the OS.
Regarding NOS/VE requiring NOS, Ken Hunter says this (in May 2007):
This is kind of true, since there was no way to shut off NOS once NOS/VE was up. Also, to access NOS/VE from anywhere but the console, you needed NOS running because unit record and interactive I/O was done through NOS. I don't think NOS/VE ever required NOS for disk I/O and paging, like Windows 3.x used DOS for this. It was also possible to boot NOS/VE from NOS/BE, and have unit record and interactive I/O go through NOS/BE. I've actually seen this done (but not at MSU). I've also seen NOS/BE hang after NOS/VE was boot, and NOS/VE kept running fine (sort of -- I think I/O going through NOS/BE stopped). In theory, it should have been possible to boot NOS/VE from any 170 state OS, such as SCOPE/Hustler, but I don't think things ever progressed that far at MSU. Dave Bright claims otherwise. If so, it was something he did which died when he left. I don't remember ever seeing this run.
The problem was that the NOS/VE bootstrap PP for NOS/BE depended on and modified NOS/BE disk tables, which had a different format in SCOPE/Hustler.