Space Shuttle IT: try landing it and launching again

News from Nasa: one of the Shuttle's four computers has gone into digital dreamland and its memory is being downloaded to Mission Control for expert picking-over. According to Houston on Thursday:

"Atlantis Commander Chris Ferguson has begun troubleshooting General Purpose Computer (GPC) 4, one of the four primary shuttle GPCs. GPC 4 was being used as the systems management computer when it failed at 6:07 p.m. EDT Thursday, generating an alarm that awakened the Atlantis crew. Ferguson configured GPC 2 as the systems management computer before going back to sleep 45 minutes later. A fifth GPC is loaded with backup software and operates independently, ready to take over in case of an endemic problem with the four primary computers.

While troubleshooting is underway on GPC 4, GPCs 1 & 2 are operating Atlantis’ systems and GPC 3 is in a standby mode. The first step in today’s procedure is underway, transmitting the memory of GPC 4 to Mission Control for evaluation."

The computers, IBM AP-101S, are a pre-microprocessor 1970s design, albeit with a late 80s refresh, and have had various glitches in the past: Nasa's Office Of Logic Design has a lot of fascinating detail about how the orbiter's digital systems work and, on occasion, don't.

This particular problem is being investigated but, as our colleagues in CBS News point out, "The troubleshooting plan for GPC-4 called for a data dump and an initial program load, or IPL, to reload the machine with software from a mass memory unit to see if it would start back up and run normally. If so, and if no other issues develop, the fault that took the computer down Thursday will be considered a transient glitch, defined as a problem that resulted in a single failure in a computer that subsequently responded normally to an IPL". In other words — try switching it off and on again.

Sound familiar? If you're in the habit of listening to mission control audio you may well have heard other such incidents. Both the International Space Station and the Shuttle have their fair share of IT failures - and as the crews aren't IT specialists, and the equipment they use these days is largely off-the-shelf stuff, the diagnostic conversations are exactly the same as the ones earthlings have. In the past, I've heard half-hour discussions about RS-232 gender issues (never resolved), a multi-session ground-to-air debug of Microsoft Outlook (if anyone should be using email in the cloud...), and more disturbingly familiar networking problems than anyone should suffer in the 21st century.

It's a long way from the inspirational heroics of Apollo 13, or the continued heroic engineering that's kept the often-faulty Voyager space probe twins in action since 1977. Yet it's only when living in space has the same rather dull certainties as life down here that we'll know we've really made it. Expect the Martian colonists, the Pluto explorers and the first mission to Betelgeuse to have to check their warranty information and listen to music on hold too.