Linked by Thom Holwerda on Mon 4th Mar 2013 20:32 UTC, submitted by Anonymous
Linux In a recent EE Times 2013 Embedded Market study, Android was the OS of choice for future embedded projects among 16 percent of the survey's participants, second only to 'in-house/custom' (at 28 percent). But if a spectrum of disparate approaches can be lumped together as a single option, why not aggregate the various shades of Linux to see how they compare? "Parsing the EE Times data this way makes it abundantly clear that Linux truly dominates the embedded market."
Thread beginning with comment 554262
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[3]: Naturally
by Alfman on Tue 5th Mar 2013 19:31 UTC in reply to "RE[2]: Naturally"
Alfman
Member since:
2011-01-28

ricegf,

"BSDs have great security, but do they handle hard reat-time"

I honestly wouldn't know, I've done several searches just now to find out and there do seem to be some who are using BSDs for realtime applications, FSMLabs comes up very prominently in those searches but they're mostly marketing pieces and their website is useless.

So like you, I'm not sure. Never the less, not everyone in the embedded space needs hard real time scheduling and I'd be curious how many devices (Linux or otherwise) in the EE study are actually hard realtime scheduled. I would think that for the majority of consumer needs, soft realtime scheduling is already good enough. Although I'm curious what it is you do that needs it.

Reply Parent Score: 3

RE[4]: Naturally
by ricegf on Wed 6th Mar 2013 11:09 in reply to "RE[3]: Naturally"
ricegf Member since:
2007-04-25

Avionics. If we missed a deadline, a computer crash would be the least of our worries. ;)

Reply Parent Score: 4

RE[5]: Naturally
by Alfman on Wed 6th Mar 2013 16:15 in reply to "RE[4]: Naturally"
Alfman Member since:
2011-01-28

ricegf,


"Avionics. If we missed a deadline, a computer crash would be the least of our worries"

I know your half joking, but my initial impression would be that realtime flight control systems are far less demanding (in terms of schedulers) than even a basic audio daemon doing realtime effects. The difference being that with an audio application, we can wave our hands and merely assume the upper bound exists for priority threads (and probably be right if distortions never happen under full load), but with important control systems someone needs to compute and actually prove a deterministic upper bound.


What is the sample/control rate used by avionic controllers?

What kind of systems would start failing on an aircraft if an operation happened 1ms late? Even the fastest servos are pretty slow in computer terms.

Do you use multi-process, multi-threaded code with multiple threads running on the same core such that the OS scheduler plays a significant role?

Feel free to educate me since I have no experience in the matter, but I am curious.

Reply Parent Score: 2