Hi Ken,
I am not sure what you mean by the products being deterministic - that is
good thing that they are. But when you mention real time performance
and Java then I think I know what you mean. Java is not deterministic in
respect to real time performance issues since it has incremental garbage
collection; you never know the Java VM stops your thread to do some
garbage collection.
I am afraid I can not help you on your second question on whether or not
the standard Java products is going to improve on real time issues. I know that Real
Time Java tries to open op the Java VM allowing application programmers to
give more precise priorities to threads (wether it is a thread to be run
real time, exact time and some more I can not remember). Also allowing you
to specify the garbage collection algoritm. And some more I can not
remember - look it up if you are interested
http://www.rtj.org/.
Best regards
Kristian
On Fri, 07 Nov 2003 18:51:38 -0800, Ken wrote:
Hi.
I'm working on a real time software project. It's not embedded, but
it certainly is considered at a minimum "soft" real time, since it
results in aircraft being moved around.
Some folks here are looking into using Java, JMS, JMX and J2EE. I'm
concerned about this because I've heard there are problems with
implementations of these standards being deterministic.
Could someone explain if determinism is a problem with this product
family, and if so, what the issues are?
I'm assuming there is a problem, thus the work being done on Real Time
Java--I just don't have a clear understanding of what the issues are.
I guess a follow up question is does anyone know if determinism issues
are being addressed for JMS, JMX or J2EE, so the could be used once
Real Time Java implementations are readily available?
Thanks so much!
Ken
______________