[extropy-chat] Re: SPACE: where are we?

[Eugen Leitl]

On Sun, Feb 08, 2004 at 09:16:44PM -0500, Dan Clemmensen wrote:

> In this case NASA was using variants of commercial hardwrae and software.

Commercially available, yes. Commodity, no. The components were military
versions (rad-hardened).

> Specifically, they were using a PPC variant CPU and VxWorks as the operating

Pretty good choice, given that the hardware choice is frozen early in mission
design. The reason the hardware looks dated is not just that military
versions are tagging a few generations behind -- the design *is* dated.

> system. The bug occurred when they filled the FLASH disk. This problem 
> is not
> unique to NASA's hardware. Use of an open-source OS would therefore have

Which open source OS? Microkernel, hard realtime, runtime patching? There is
no such OS.

> been useful. Among other things, if they had been using an open embedded OS,

Which one?

> they could have published all the code and asked for volunteer 
> reviewers. You cannot

They would have to release the emulator as well. Most open source people are
unfamiliar with deep embedded development. They can't code for it, they can't
debug for it. Only very few of them can at all produce code capable of
passing harsh aerospace quality control.

> easily do this with VxWorks code, because the volunteers would need VxWorks
> licenses to run the code in emulation.

It's too bad we don't have a hard realtime kernel that is open sourced.
Maybe, one day, and a mil-spec open core to match. Meanwhile, I'm not holding
my breath.
 
> There are several viable embedded OSes. Linux is quite suitable when you 

Linux is extremely unsuitable for aerospace hard realtime embeddeds.

> have as much
> Flash and RAM as NASA has available (512MB and 256MB respectively) Linux can

128 MBytes flash.

> run easily on a Palm pilot wiht 8MB, and can be cut down much further if 
> needed.

Most of the code onboard isn't the OS. That spacecraft have a lot of other
things to do than to run desktopcentric bloated monolithic operating systems.

-- Eugen* Leitl <a href="http://leitl.org">leitl</a>
______________________________________________________________
ICBM: 48.07078, 11.61144            http://www.leitl.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE
http://moleculardevices.org         http://nanomachines.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20040209/efdf47dd/attachment.bin>