Magnetic force

I have an optically-couple "supervisory port" on my
devices.  A design requirement is that the outer surfaces
of the devices must be "wipe clean" -- no ridges or
grooves.
I mechanically attach the mating "cable" to the port
magnetically (an idea I stole from designing electric
power meters -- but, they aren't "wipe clean"!)
As I can't rely on any significant mechanical structures
to help maintain the attachment's position, that must
be guaranteed solely by the strength of the magnetic
bond.
I have a 1/4" dia glass "jewel" that is just barely convex
to protect the optical port and support light passage.
The slight bump helps locate the port as well as some small
assistance in keeping the mated "cable" positioned properly.
In no case can the needs of the port exceed a concentric
diameter of 1/2".  I plan for the magnet to only be present
in the mating cable so its adhesion will be determined by
the metallic ring surrounding the jewel.
I figure I need to specify a magnet strong enough to
"support" (against the force of gravity exerted on the
mass of the cable) the cable in a horizontal position
(i.e., "mated to the ceiling").  And, to support the
cantilevered weight of the cable when mounted
vertically.
I may opt to implement the "cable" as a wireless dongle
if the weight of the cable assembly starts to compromise that
magnetic adherence.
I guess I also need to consider the magnetic force exerted
through the friction between the mating surfaces when
mounted vertically (so the cable doesn't "slide down the wall")
Anything else I should consider?  I realize that the cable end
now will have a tendency to want to grab onto metallic objects
so have to factor that into my dongle vs. cable decision
(it's a lot easier to control the position of a dongle in
space than the end of a possibly *dangling* cable!)
I suspect the adhering "metal" piece has to be on the skin
of the device as the magnetic field falls off quickly with
distance -- diluting the ability of the cable to remain
EXACTLY where desired.  A "decal" over the surface likely
would be tolerable...