Re: What's Real-Time? (was Re: Widespread C++ Competency Gap?)

[dgy@rtd.com (Don Yuniskis)]

In article <hbaker-3012941712040001@192.0.2.1>,
Henry Baker <hbaker@netcom.com> wrote:

>In the RT community there are the terms '_hard_ real-time' and
>'_soft_ real-time'.  'Hard' RT means that the consequences of missing
>a deadline are catastrophic, while 'soft' RT means that the consequences
>are expensive, but not catastrophic.

Yes, as I indicated in my reference to the value function associated
with the missed deadline (for t>deadline, value <= 0 for HRT; value > 0
for SRT).

>Designing for a 'hard' RT system is different from designing for a
>'soft' RT system because you have to find the latencies of _every_
>operation, no matter how rare -- e.g., you have to worry about the
>latencies of rare combinations of rare events -- e.g., multiple page
>faults from a single instruction, cache faults for every memory reference
>of a 1000-instruction sequence, etc.

Exactly!  A fact that most designers neglect either from ignorance or
by choice (it's orders of magnitude harder to design for the worst-case
whether you're designing hardware OR software -- tho', admittedly, a
worst-case hardware design is orders of magnitude EASIER than a
worst-case software design ;-) :-(

>This means that nearly everyone is looking for ways to convert 'hard'
>RT systems into 'soft' RT systems so that they can take advantage of
>the increased average speeds.
>
>So if you're still doin' 'hard time', wise up!  :-) :-)

Right.  However, certain parts of evry application "must" be HRT.
The trick is to move as much of the application into the domain of
SRT -- "Gee, it would be real nice if this got done on time... but
it's more important that I know how to deal with the consequences
of when it DOESN'T get done on time!"