PASADENA, Calif. – NASA's Galileo spacecraft began what would become a 14-year odyssey of exploration 20 years ago this Sunday, Oct. 18. Galileo was humanity's first emissary to orbit a planet in the outer solar system – Jupiter. (Read More)
20 years ago I was just starting to work with a remarkably small, very unix-like kernel from this *tiny* company in Emeryville (near Berkeley). You could run it with a 68020SBC and a single 2MB memory board, a computer costing perhaps only twenty thousand dollars. It was incredibly fast (considering the hardware) and even supported networking. It was called VxWorks, and it just came out with a "Wind" kernel. Back then, no-one was talking much about sending commercial off-the-shelf stuff into space.
Two years later I was working at Wind River, something like Employee #79 or so, the company had moved to Alameda and was growing. The space industry used us in their machine test-beds… but not run the space probes. It was a dream to be used in a project that would actually leave the planet.
The first time I visited JPL, Galileo was in (I think) the High-Bay. You could go in this little observation room and look at it. It was about the size of a small sea-freight container. A true Titan, as far as deep-space craft go. Consider that most of the craft I've worked with are smaller than Volkswagens, this was huge by comparison. I'll always think of Galileo as "the last of the Titans" for this reason. "They don't build them like that any more!"
Since then most of the missions I've seen used the "COTS" method and shoestring budgets, re-using as much from the previous missions as possible, and making things as small as possible. Some of these satellites were perhaps slightly larger than basketballs, some deep space missions were as large as a large go-cart (including bumpers). But none are near so large as Galileo.
When the Space Agencies were first making these changes – moving to COTS hardware and software parts, it was a huge gamble. No-one knew for sure if the commercial offerings were up to the challenges NASA had planned. NASA has made part of it's fame on the "Failure Is Not An Option" work ethic, could COTS actually work out?
The proof is in the pudding now. In evidence we have some of the most successful deep space and planetary probe missions ever attempted, still running, some cases having fulfilled their missions and continued for more than another 20-mission-lifetimes.
NASA has managed to switch, from it's legacy of titanic space craft, to a new generation of lean robotic craft. Here's to a continued legacy of excellence, from a new breed of smaller craft.
The following cast of characters, and more, all run, or ran, flavors of VxWorks: Pathfinder, GP-B, SORCE, MESSENGER, DS-1, Stardust, MO 2001, MER 2003, MRO, Phoenix, LRO, LCROSS, Spitzer, STEREO, Kepler, Rosetta, Genesis, Deep Impact, DIXI/EPOCH, NEXT, and the Deep Space Network of Antennae.