This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: How does people think of ecos's future compare to so m any existing RTOS? eg.RT-Linux, Embedded Linux, VxWorks, etc.. Welcome t o discuss


I believe it is far more important to understand the underlying concepts
behind an RTOS than the implementation of a specific RTOS itself, as it is
essential to understand these concepts to write effective applications for
them. I'm still amazed at the lack of understanding, often by very
experienced engineers, about the benefits of using an RTOS in event-driven
systems and the perception that they add unnecessary overhead; I can imagine
this is one of the main reasons for over 50% (AFAIK) of so-called embedded
OSes still being 'home-grown', i.e. they are probably not really 'proper'
OSes at all, but a simple non-preemptive scheduler, which schedules tasks
that are fixed and specific to their own system.

Once the concepts are understood, which RTOS is used is really irrelevant as
I think it is fair to say that all embedded RTOSes offer equivalent
services. So the choice is basically down to preference more than anything.
However, I would say Linux (which was originally mentioned), which is now a
serious contender as an embedded OS, is more complex. I think it takes more
than, say, 3 days to get your head around user-mode process-based
applications (with such concepts as forking and signalling), user-mode
thread-based applications (more familiar to RTOS people but still with
caveats), kernel mode applications etc. If you then mix RT-Linux or RTAI
into the pot, you are talking about a pretty complex situation to fully get
to grips with. I have also found getting documentation for Linux programming
a bit fraught as it is scattered over the Internet (often in mailing lists)
and the quality is variable. IMHO eCos is certainly easy to get to grips
with from the programming model, the source and the documentation. The
documentation is comprehensive and considerably better than many commercial
OSes I have come across.

I think it is worth considering what sort of system eCos is targeted at. It
seems to me that eCos does overlap considerably with embedded Linux for
embedded 32 bit processor designs and that embedded Linux may be preferred
simply because it is 'trendier' than eCos and there has been more stuff
written for Linux. For example, the Sharp Zaurus puts a big emphasis on the
fact that it is Linux-based, and the Trolltech Qtopia windowing system was
leveraged from Trolltech's desktop Qt system. I doubt eCos (at the moment)
would be a strong selling point, plus (AFAIK) Microwindows doesn't really
have any supporting apps. If you consider the target market for eCos is
smaller, resource limited embedded systems, then I guess it is preferable to
Linux, however it would be nice if eCos could run on 8 bit micros like PIC,
8051 etc. Many small embedded systems still use 8 bit micros and new SoC
designs are appearing with 8 bit micros in them, as they offer small silicon
real estate, low power etc. This would then extend eCos in an area where it
is, IMHO, innately suitable due to its scalability and low resource
requirements. Could this be done, or is the lack of gcc support an issue?

Robert Cragie, Design Engineer
_______________________________________________________________
Jennic Ltd, Furnival Street, Sheffield, S1 4QT,  UK
http://www.jennic.com  Tel: +44 (0) 114 281 2655


> -----Original Message-----
> From: ecos-discuss-owner@sources.redhat.com
> [mailto:ecos-discuss-owner@sources.redhat.com]On Behalf Of NavEcos
> Sent: 01 October 2002 22:13
> To: Doug Fraser; Qiang Huang; Ecos-Discuss
> Subject: Re: [ECOS] How does people think of ecos's future compare to so
> m any existing RTOS? eg.RT-Linux, Embedded Linux, VxWorks, etc.. Welcome
> t o discuss
>
>
> On Tuesday 01 October 2002 08:02 am, Doug Fraser wrote:
> > I have been doing EMBEDDED development for eighteen years,
> > small scale and large scale, using anything from 'roll your own'
> > on 16 bit DSPs and MPUs to pSOS, vxWorks, Precise and QNX on x86,
> > 680x0, ARM, (a little MIPs), MPC8xx, and MPC750.
> >
> > Price and 'will it do the job' have certainly been at the fore
> > in decisions about the software development environment.
> > But ramp up and the existing knowledge pool among the available
> > headcount as well as new hires has always been part of the
> > analysis. Education and training for an unfamiliar OS cost hard
> > dollars as well as time to market.
> >
> > I am well aware that pennies count. I have worked on modems.
> > If there is a commodity market, that is it.
> >
> > That being said, I am a fan of eCos. Otherwise I wouldn't
> > bother making a reply. For small, deeply embedded systems,
> > eCos packs a lot of capability into a very small package.
>
> You have me beat by 8 years, but I cannot think of a single company
> that would worry about previous experience with an OS to chose an OS.
> It is trivial to learn a new embedded OS and if I had an engineer
> under me that couldn't do it, I'd fire him on the spot for gross
> incompetence to be really blunt about it, at least in this job
> market.
>
> All of the companies I have worked for, except Microsoft (for obvious
> reasons), would switch embedded OSes on the fly as needed or
> dictated or needed depending on who made the decision to switch,
> engineering or management and it was simply expected the engineering
> team could adapt.  The APIs between most embedded OSes are nearly
> identical.
>
> > Doug Fraser
> >
> > > -----Original Message-----
> > > From: NavEcos [mailto:ecos@navosha.com]
> > > Sent: Tuesday, October 01, 2002 9:51 AM
> > > To: Doug Fraser; Qiang Huang; Ecos-Discuss
> > > Subject: Re: [ECOS] How does people think of ecos's future
> > > compare to so
> > > m any existing RTOS? eg.RT-Linux, Embedded Linux, VxWorks,
> > > etc.. Welcome
> > > to discuss
> > >
> > > It has never been a consideration at any company I worked with for
> > > EMBEDDED operating systems what the engineers already know.  I have
> > > worked with pSOS, vxWorks, WinCE, BlueCat, and several roll-your-own
> > > OSes.  I think in the majority of corporations all that matters for
> > > an embedded OS is first, can it do the job (this includes tools),
> > > and second, what does it cost.  In embedded applications, pennies
> > > count.
> > >
> > > With Linux and Windows, there is a question of administration and a
> > > pre-installed userbase that doesn't exist with embedded operating
> > > systems.
> > >
> > > > Doug Fraser
> > > >
> > > > > -----Original Message-----
> > > > > From: NavEcos [mailto:ecos@navosha.com]
> > > > > Sent: Tuesday, October 01, 2002 5:46 AM
> > > > > To: Qiang Huang; Ecos-Discuss
> > > > > Subject: Re: [ECOS] How does people think of ecos's future
> > > > > compare to so
> > > > > many existing RTOS? eg.RT-Linux, Embedded Linux, VxWorks,
> > > > > etc.. Welcome
> > > > > to discuss
> > > > >
> > > > > On Tuesday 01 October 2002 02:14 am, Qiang Huang wrote:
> > > > > > Just curious about the future of ecos compare to other
> > > > >
> > > > > RTOS, anybody have
> > > > >
> > > > > > any idea?
> > > > > >
> > > > > > thanks
> > > > >
> > > > > I think that if Linux is to Microsoft Windows, that eCos is
> > > > > to vxWorks.
> > > > >
> > > > > In my opinion, companies are CRAZY to be willing to pay $100K
> > > > > plus royalties
> > > > > to develop with vxWorks when eCos exists.  For some
> > > > > applications you need the
> > > > > greater complexity of vxWorks, but not many of them.
> > > > >
> > > > > -Rich
> > > > >
> > > > > --
> > > > > Before posting, please read the FAQ:
> > > > > http://sources.redhat.com/fom/ecos
> > > > > and search the list archive:
> > >
> > > http://sources.redhat.com/ml/ecos-discuss
>
>
> --
> Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss
>
>


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]