The ads implied the device worked on jacketed fibres; I suppose that depends on the transparency of the jacket to 1064nm IR radiation.
It seems like they learned about a neat concept, made a proof of concept that actually worked, but ran into physical reality when they tried to turn it into a reliable product (while being optimistic along the way). It's real hard to stop being optimistic at exactly the right time, and I'd be pretty forgiving if it seemed like everybody was giving an honest effort.
If I was to build a queryable database, I'd grab something like Lucene, and stick that on top of something like Postgres. The next step would be to build an interface so my non-technical business partners don't have to write SQL or some other query and manually send that to the server, but instead can ask stupid questions like "how long are Trumps fingers" and get results. That last part is PRISM.