I cannot say I really don't know but here's what I see:
I'm not sure that your NAS allows two logins
I see two possibilities:
there would be a conflict in the kernel if the drive were mounted 2x of course (no way to know which update was "better" where a conflicting update was requested), and the kernel wouldn't allow it anyway, it would be non-atomic and end in corruption despite best efforts and guesses.
It seems to me you might try connecting to the NAS drive as a different user (ie, having two users active at same time on your desktop, ie via su(1)).
that is: you have a "network attached storage" and a single (login) to it. on that mount you can only access as the user you attached as (how apple does any mounted device). (you'd have to connect as the user who can read all parts of the drive you want read)
to read the device, user level authority would then exist (permissions on files) once connected.
you might then have "keyed access to that storage", perhaps one per user - if the driver for the device does not support file permissions and needs an interface to ask for keys. but i doubt this case happens.
----------------------------
possibility two is: your NAS does not allow remote mounting because it does all atomic (wheel) updating and does not allow external OS control, because multiple connections of different OS running in different space and time would certainly corrupt data. (in other words, it has an internal OS, and is basically like connecting to remote NFS)
the NAS is never "mounted" as a disk by Apple (that while your shown an icon, it is not a mount icon), and two different users read the NAS separately
in that case each session to accessing should be by a driver which has opened a session. i imagine such a driver would allow at least one per user concurrently if not multiple per user, and cannot say
------------------------------
i would be suspicious of any software that allowed you multiple connections if they really were separately mounted connections. OS run in different time and space. there is no known way to synchronize two of them because conflict management always must fail in more complicated cases.
perhaps what you've seen is an OS which is not doing what you think it is with the connection (it is bypassing security or worse: is using a shaky algorithm that does not really keep concurent wheel requests synchronized in the end)
