Re: [rtl] IP useage from rtlinux, reply to Richard Teltz

[Richard Teltz <rich@immrc.eng.mcmaster.ca>]

Warren,

Somewhere in your message there was a question..."What do I
hope to gain...?"

The goal of the system we have developed (VxWorks, VME, etc)
is to provide an *open* architecture for industrial automation
systems.  We chose IP because it is almost universally supported.
You would be surprised what kind of performance even a 10Mb 
ethernet can provide within a small subnet typical of a 
manufacturing cell.   Deterministic it is not, fast it can be.

ARCNET, CANbus, etc are not supported by a terrifically wide number of
vendors, and QNX is great if all your machines are QNX (kind of like
MicroSoft).  

UDP is a nice option because you can build mechanisms on top of the
protocol to check timing, etc.  At least you know when you do not meet
deadlines and can react accordingly.  In fact someone at Stanford did a
nice job of this and is now selling the product through Real Time
Innovations, a VxWorks VAR.  This is not widely supported either however,
but you can still interact with other systems without placing such
requirements on them (a somewhat "open" solution).  

While textbook discussions on determinism have their value,
in reality a RTOS still only provides you with the ability to guarantee
one thing at any one time.  Arguably, you could do this with a NMI
on just about any machine with any OS.  Aa a result, a lot of effort goes
into designing RT systems so that the "one time" part of this does not
conflict for different entities (I am assuming a uniprocessor system). 

However, RTOS's do offer a lot of other handy things, not the least
of which is (usually) a simple and efficient scheduling policy with a
simple API.  This is mainly what I am after in my application, in addition
to avoiding the big bucks required for VxWorks, QNX, or others.

Rich

*------------------------------------------------------------------*
|  INTELLIGENT MACHINES AND MANUFACTURING RESEARCH CENTRE (IMMRC)  |
*------------------------------------------------------------------*
|       Richard Teltz,           | Dept. of Mechanical Engineering,|  
|  principal research engineer,  | McMaster University,            |
|   tel: (905) 525-9140, x27351  | Hamilton, Ontario,              |
|   fax: (905) 572-7944          | Canada,                         |
| email: rich@immrc.mcmaster.ca  | L8S 4L7                         |
*------------------------------------------------------------------*


--- [rtl] ---
For more information on Real-Time Linux see:
http://www.rtlinux.org/