vibration testing introduction

hello,

i'm a mechanical engineer and with my co-workers, we're trying to revamp our company's vibration testing criteria. we make electrionic fixtures which range from rackmount equipment to lighting fixtures using solid state lighting.

i understand the obvious differences between random and sinusoidal. but my confusion lies in how to test for a particular application. most of our products are installed in a static location, bolted to a junction box, or bolted to a steel structure like a bridge.

when testing for this type of application, does it make sense to specify a sinusoidal or random vibration test by a g value or PSD? or is it more practical to specify an amplitude of displacement, direction of displacement, and frequency of displacement?

also, can anyone point me in the direction of some good text books or reading material to help us on our way?

thank you.
Steve Kondo