I'm mostly using cellular/mobile vehicle trackers (of which there are hundreds of vendors), but for regions where there's no cell/mobile coverage, we fall back to satellite options. With lots of optimization, it's only slightly more expensive for our use-case.
Specifically, we deploy these units:
http://www.skywave.com/en/our-technology/satellite-cellular-...
which use the Inmarsat constellation. The specs probably aren't quite up to being attached directly to an airplane, but with the right housing and a heater, they're probably not far off.
The trick for finding robust off-the-shelf gear is to look for industries that operate in remote regions, like oil and gas, and maritime. They also have fairly harsh operating conditions, so equipment sold there is usually fairly tough.
Prices have really come down over the last 5 years.
What I don't get is why airlines insist on bundling emergency telematics along with position tracking. It massively complicates what otherwise is a pretty simple problem. It feels like that description of the Bradley Fighting Vehicle out of the Pentagon Wars movie.
With a box the size of a Big Mac, at <$1k per unit and about $100/month in data costs, you could deploy an off-the-shelf industrial grade tracking unit that would transmit location/speed/altitude every 10 seconds whenever the unit is powered. All it would need would be a power supply. This would neatly solve the "where the heck is the plane" problem, globally, for a fairly trivial cost.
I realize that finding a spot on the plane to mount it, verifying aerodynamics, figuring out how to wire it in, testing, doing the work across a fleet and managing all that data all end up costing a lot more than the equipment, but given the costs being quoted for some solutions in the articles that have been coming up, it really seems like the airlines could have a big PR win by paying for it with petty cash out of their marketing/advertising budget.